#SampleID	BarcodeSequence	LinkerPrimerSequence	LAB_PERSON_CONTACT	TARGET_SUBFRAGMENT	ASSIGNED_FROM_GEO	RUN_DATE	TITLE	AGE	INVESTIGATION_TYPE	HAND_KEY	INCLUDES_TIMESERIES	BODY_SITE	POOL_MEMBER_NAME	ELEVATION	COLLECTION_DATE	ALTITUDE	SEX	HOST_SUBJECT_ID	ANONYMIZED_NAME	POOL_PROPORTION	SAMP_SIZE	PRINCIPAL_INVESTIGATOR	STUDY_DESCRIPTION	LONGITUDE	HAND	STUDY_CENTER	BODY_HABITAT	BARCODE_READ_GROUP_TAG	STUDY_ABSTRACT	ENV_FEATURE	KEY_SEQ	POOL_MEMBER_ACCESSION	BODY_PRODUCT	REGION	PCR_PRIMERS	KEY_SIZE	OWNER	PMID	PUBLIC	LAB_PERSON	EXPERIMENT_CENTER	COUNTRY	COMMON_NAME	DEPTH	HOST_TAXID	TAXON_ID	SUBMIT_TO_INSDC	USE_FREQUENCY	PROJECT_NAME	SEQUENCING_METH	EXPERIMENT_ALIAS	RUN_PREFIX	ENV_BIOME	TYPING_DIGIT	PLATFORM	RUN_ALIAS	STUDY_TITLE	STUDY_ALIAS	EXPERIMENT_TITLE	SAMPLE_CENTER	SPACEBAR_OWNER_SEX	AGE_UNIT	STUDY_ID	LIBRARY_CONSTRUCTION_PROTOCOL	STUDY_REF	HOST_COMMON_NAME	MIENS_COMPLIANT	ENV_MATTER	TARGET_GENE	RUN_CENTER	PRIMER_READ_GROUP_TAG	AGE_IN_YEARS	INSTRUMENT_NAME	LATITUDE	EXPERIMENT_DESIGN_DESCRIPTION	Description
M9Gkey217.141099	AGAGCAAGAGCA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Gkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Gkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Gkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Gkey
M3Space217.141098	ATCGATCTGTGG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	Mixed	0	NA	M3Space217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M3Qkey217.141051	AGTACTGCAGGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Qkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Qkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Qkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Qkey
M9Rinr217.141080	ACTCACGGTATG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	right	0	UBERON:skin of finger	M9Rinr217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Rinr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M9Rinr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right ring finger tip
M3Xkey217.141055	ATAGCTCCATAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Xkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Xkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Xkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Xkey
M2Thml217.141005	AACGCACGCTAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	left	0	UBERON:skin of finger	M2Thml217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Thml217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M2Thml217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left thumb tip
R1Space217.141077	ATCTCTGGCATA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	R1Space217_V2	1624.0	7/15/2008	0.0	NA	232:R1	R1Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	R1Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	public	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	mixed	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M9Ckey217.141013	AGCACACCTACA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Ckey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Ckey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Ckey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ckey
M9Midl217.141043	ACGTGCCGTAGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	left	0	UBERON:skin of finger	M9Midl217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Midl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M9Midl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left middle finger tip
M9Midr217.141060	ACTATTGTCACG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	right	0	UBERON:skin of finger	M9Midr217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Midr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M9Midr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right middle finger tip
M2Zkey217.141015	ACGACGTCTTAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Zkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Zkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Zkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Zkey
M9Rinl217.141020	ACGTTAGCACAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	left	0	UBERON:skin of finger	M9Rinl217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Rinl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M9Rinl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left ring finger tip
M2Gkey217.141106	ACCAGACGATGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Gkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Gkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Gkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Gkey
M9Pkey217.141096	AGAACACGTCTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Pkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Pkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Pkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Pkey
M2Space217.140997	ACGTACTCAGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	Mixed	0	NA	M2Space217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Indl217.141074	AACTCGTCGATG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	left	0	UBERON:skin of finger	M2Indl217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Indl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M2Indl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left index finger tip
M3Indl217.141067	AGCTATCCACGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	left	0	UBERON:skin of finger	M3Indl217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Indl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M3Indl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left index finger tip
M3Lkey217.141097	ATACTATTGCGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Lkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Lkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Lkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Lkey
M2Wkey217.141070	ACACTAGATCCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Wkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Wkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Wkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Wkey
M9Akey217.141086	AGACCGTCAGAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Akey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Akey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Akey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Akey
M3Lsft217.141088	ATCAGGCGTGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Lsft217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Lsft217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Lsft217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	medium	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Left_Shift
M3Jkey217.141017	ATACAGAGCTCC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Jkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Jkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Jkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Jkey
M3Ekey217.141042	AGTCACATCACT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Ekey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Ekey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Ekey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ekey
M2Lsft217.141056	ACGGATCGTCAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Lsft217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Lsft217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Lsft217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	medium	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Left_shift
L3Space217.141024	ATGCGTAGTGCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	L3Space217_V2	1624.0	7/15/2008	0.0	NA	232:L3	L3Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	L3Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	public	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	mixed	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
F12Space217.141068	ATCTGGTGCTAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	F12Space217_V2	1624.0	7/15/2008	0.0	NA	232:F1	F12Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	F12Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	female	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M3Indr217.140993	AGGACGCACTGT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	right	0	UBERON:skin of finger	M3Indr217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Indr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M3Indr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right index finger tip
M3Hkey217.141093	ATACACGTGGCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Hkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Hkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Hkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Hkey
M9Xkey217.141072	AGATGTTCTGCT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Xkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Xkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Xkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Xkey
M10Space217.141036	ATCGCGGACGAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	M10Space217_V2	1624.0	7/15/2008	0.0	male	232:M1	M10Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M10Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Indr217.141062	AATCGTGACTCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	right	0	UBERON:skin of finger	M2Indr217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Indr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M2Indr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right index finger tip
M9Enter217.141084	AGCGAGCTATCT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Enter217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Enter217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Enter217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	medium	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ente
M9Qkey217.141019	ACTCGATTCGAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Qkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Qkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Qkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Qkey
M11Space217.141008	ATCGTACAACTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	M11Space217_V2	1624.0	7/15/2008	0.0	male	232:M1	M11Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M11Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
F11Space217.141071	ATCTACTACACG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	F11Space217_V2	1624.0	7/15/2008	0.0	female	232:F1	F11Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	F11Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	female	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Rinl217.141058	AAGAGATGTCGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	left	0	UBERON:skin of finger	M2Rinl217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Rinl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M2Rinl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left ring finger tip
M3Vkey217.141028	ATATCGCTACTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Vkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Vkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Vkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Vkey
M2Enter217.141012	ACGGTGAGTGTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Enter217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Enter217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Enter217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	medium	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ente
M2Thmr217.141087	AATCAGTCTCGT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	right	0	UBERON:skin of finger	M2Thmr217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Thmr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M2Thmr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right thumb tip
M9Ykey217.141029	ACTGTACGCGTA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Ykey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Ykey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Ykey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ykey
M3Kkey217.141046	ATACGTCTTCGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Kkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Kkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Kkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Kkey
M9Fkey217.141065	AGAGAGCAAGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Fkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Fkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Fkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Fkey
M3Thmr217.140996	AGCTTGACAGCT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	right	0	UBERON:skin of finger	M3Thmr217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Thmr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M3Thmr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right thumb tip
M9Vkey217.141018	AGCACGAGCCTA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Vkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Vkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Vkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Vkey
M2Jkey217.141038	ACCGCAGAGTCA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Jkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Jkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Jkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Jkey
M9Pinr217.141002	ACTCAGATACTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	right	0	UBERON:skin of finger	M9Pinr217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Pinr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M9Pinr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right pinkie finger tip
M9Skey217.141004	AGACGTGCACTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Skey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Skey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Skey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Skey
M3Rinr217.141091	AGGTGTGATCGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	right	0	UBERON:skin of finger	M3Rinr217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Rinr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M3Rinr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right ring finger tip
M3Akey217.141006	AGTGTTCGATCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Akey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Akey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Akey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Akey
M2Qkey217.141102	ACACGGTGTCTA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Qkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Qkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Qkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Qkey
M9Thml217.141021	ACGTCTGTAGCA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	left	0	UBERON:skin of finger	M9Thml217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Thml217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M9Thml217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left thumb tip
M3Ykey217.141023	AGTCTCGCATAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Ykey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Ykey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Ykey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ykey
M9Kkey217.141033	AGATACACGCGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Kkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Kkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Kkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Kkey
F10Space217.141045	ATCGCTCGAGGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	F10Space217_V2	1624.0	7/15/2008	0.0	female	232:F1	F10Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	F10Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	female	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Lkey217.141048	ACCTGTCTCTCT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Lkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Lkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Lkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Lkey
M3Bkey217.141059	ATATGCCAGTGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Bkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Bkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Bkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Bkey
M2Ckey217.141092	ACGATGCGACCA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Ckey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Ckey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Ckey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ckey
M2Rinr217.141027	ACACATGTCTAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	right	0	UBERON:skin of finger	M2Rinr217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Rinr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M2Rinr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right ring finger tip
M3Midl217.141103	AGCTCCATACAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	left	0	UBERON:skin of finger	M3Midl217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Midl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M3Midl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left middle finger tip
M3Pinl217.141094	AGCTGACTAGTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	left	0	UBERON:skin of finger	M3Pinl217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Pinl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M3Pinl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left pinkie finger tip
M2Ukey217.141085	ACAGCTAGCTTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Ukey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Ukey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Ukey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ukey
M2Tkey217.141054	ACAGAGTCGGCT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Tkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Tkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Tkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Tkey
M9Hkey217.140999	AGAGTAGCTAAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Hkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Hkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Hkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Hkey
M3Mkey217.141107	ATCACTAGTCAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Mkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Mkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Mkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Mkey
M9Okey217.141076	ACTTGTAGCAGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Okey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Okey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Okey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Okey
M3Zkey217.141083	ATACTCACTCAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Zkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Zkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Zkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Zkey
M9Bkey217.141032	AGCAGCACTTGT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Bkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Bkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Bkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Bkey
M2Pkey217.141105	ACATCACTTAGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Pkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Pkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Pkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Pkey
M2Nkey217.141052	ACGCGCAGATAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Nkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Nkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Nkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Nkey
M3Pinr217.141057	AGTACGCTCGAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	right	0	UBERON:skin of finger	M3Pinr217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Pinr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M3Pinr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right pinkie finger tip
M3Midr217.141101	AGGCTACACGAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	right	0	UBERON:skin of finger	M3Midr217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Midr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M3Midr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right middle finger tip
M3Tkey217.141081	AGTCTACTCTGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Tkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Tkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Tkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Tkey
M3Rsft217.140995	ATCCGATCACAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Rsft217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Rsft217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Rsft217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	medium	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Right_shift
M2Ykey217.141095	ACAGCAGTGGTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Ykey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Ykey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Ykey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ykey
M9Indl217.141066	ACGTGAGAGAAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	left	0	UBERON:skin of finger	M9Indl217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Indl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M9Indl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left index finger tip
M2Vkey217.141022	ACGCAACTGCTA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Vkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Vkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Vkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Vkey
M2Akey217.141030	ACATGATCGTTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Akey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Akey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Akey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Akey
M3Nkey217.141001	ATCACGTAGCGG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Nkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Nkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Nkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Nkey
M2Skey217.141089	ACATGTCACGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Skey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Skey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Skey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Skey
M2Pinr217.141039	ACACGAGCCACA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	right	0	UBERON:skin of finger	M2Pinr217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Pinr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M2Pinr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right pinkie finger tip
M9Thmr217.141049	ACTACGTGTGGT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	right	0	UBERON:skin of finger	M9Thmr217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Thmr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M9Thmr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right thumb tip
M2Fkey217.141078	ACCACATACATC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Fkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Fkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Fkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Fkey
M9Nkey217.141010	AGCAGTCGCGAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Nkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Nkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Nkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Nkey
M9Wkey217.141090	ACTCGCACAGGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Wkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Wkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Wkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Wkey
L1Space217.141069	ATGCAGCTCAGT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	L1Space217_V2	1624.0	7/15/2008	0.0	NA	232:L1	L1Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	L1Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	public	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	mixed	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Midl217.141009	AACTGTGCGTAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	left	0	UBERON:skin of finger	M2Midl217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Midl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M2Midl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left middle finger tip
M3Wkey217.141104	AGTAGTATCCTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Wkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Wkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Wkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Wkey
M2Okey217.141061	ACAGTTGCGCGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Okey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Okey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Okey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Okey
U3Space217.141064	ATGAGACTCCAC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	U3Space217_V2	1624.0	7/15/2008	0.0	NA	232:U3	U3Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	U3Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	public	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	mixed	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Pinl217.141003	AAGCTGCAGTCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	left	0	UBERON:skin of finger	M2Pinl217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Pinl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M2Pinl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left pinkie finger tip
M9Space217.141079	AGCGCTGATGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	Mixed	0	NA	M9Space217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M3Rkey217.141025	AGTCCATAGCTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Rkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Rkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Rkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Rkey
M3Gkey217.141037	ATAATCTCGTCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Gkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Gkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Gkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Gkey
M3Ckey217.141073	ATAGGCGATCTC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M3Ckey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Ckey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Ckey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ckey
M2Bkey217.141063	ACGCGATACTGG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Bkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Bkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Bkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Bkey
M3Rinl217.141100	AGCTCTCAGAGG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	left	0	UBERON:skin of finger	M3Rinl217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Rinl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M3Rinl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left ring finger tip
M2Ekey217.141011	ACACTGTTCATG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Ekey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Ekey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Ekey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ekey
M9Ekey217.141044	ACTCTTCTAGAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Ekey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Ekey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Ekey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	7plus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ekey
M3Thml217.141040	AGCGTAGGTCGT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	33.0	mimarks-survey	left	0	UBERON:skin of finger	M3Thml217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Thml217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M3Thml217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	thumb	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	33	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left thumb tip
M2Ikey217.141000	ACAGTGCTTCAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Ikey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Ikey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Ikey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Ikey
U2Space217.141047	ATGACTCATTCG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	U2Space217_V2	1624.0	7/15/2008	0.0	NA	232:U2	U2Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	U2Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	public	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	mixed	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M9Mkey217.141007	AGCATATGAGAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M9Mkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Mkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Mkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Mkey
M2Dkey217.140994	ACATTCAGCGCA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Dkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Dkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Dkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Dkey
M2Hkey217.141026	ACCAGCGACTAG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Hkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Hkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Hkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Hkey
M9Indr217.140998	ACTAGCTCCATA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	right	0	UBERON:skin of finger	M9Indr217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Indr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M9Indr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right index finger tip
M9Pinl217.141035	ACTACAGCCTAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	25.0	mimarks-survey	left	0	UBERON:skin of finger	M9Pinl217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Pinl217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	left	CCME	UBERON:skin	M9Pinl217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	25	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	left pinkie finger tip
U1Space217.141075	ATGACCATCGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	NA	0	NA	U1Space217_V2	1624.0	7/15/2008	0.0	NA	232:U1	U1Space217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	U1Space217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	large	public	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	NA	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	mixed	NA	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Space_bar
M2Kkey217.141041	ACCTCGATCAGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Kkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Kkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Kkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Kkey
M2Rsft217.141034	ACGCTCATGGAT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Rsft217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Rsft217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Rsft217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	medium	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Right_shift
M2Xkey217.141016	ACGAGTGCTATC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Xkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Xkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Xkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1minus	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	ring	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Xkey
M2Rkey217.141050	ACAGACCACTCA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M2Rkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Rkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Rkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	6to7	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Rkey
M9Dkey217.141053	AGACTGCGTACT	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	left	0	NA	M9Dkey217_V2	1624.0	7/15/2008	0.0	male	232:M9	M9Dkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M9Dkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Dkey
M3Pkey217.141031	AGTGTCACGGTG	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M3Pkey217_V2	1624.0	7/15/2008	0.0	male	232:M3	M3Pkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M3Pkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	1to2	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	pinky	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Pkey
M2Mkey217.141014	ACGCTATCTGGA	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	NA	mimarks-survey	right	0	NA	M2Mkey217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Mkey217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	NA	CCME	NA	M2Mkey217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:surface	TCAG	NA	NA	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	small	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	metagenomes	0	NA	408169	n	2to6	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:anthropogenic habitat	index	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:surface	16S rRNA	CCME	V2	NA	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	Mkey
M2Midr217.141082	ACACACTATGGC	CATGCTGCCTCCCGTAGGAGT	chris.lauber@gmail.com	V2	n	8/14/08	Forensic_identification_using_skin_bacterial_communities	36.0	mimarks-survey	right	0	UBERON:skin of finger	M2Midr217_V2	1624.0	7/15/2008	0.0	male	232:M2	M2Midr217	0.8	1, swab	Fierer	forensic_keyboard	-105.2705	right	CCME	UBERON:skin	M2Midr217_V2	Recent work has demonstrated that the diversity of skin-associated bacterial communities is far higher than previously recognized, with a high degree of interindividual variability in the composition of bacterial communities. Given that skin bacterial communities are personalized, we hypothesized that we could use the residual skin bacteria left on objects for forensic identification, matching the bacteria on the object to the skin-associated bacteria of the individual who touched the object. Here we describe a series of studies de-monstrating the validity of this approach. We show that skin-associated bacteria can be readily recovered from surfaces (including single computer keys and computer mice) and that the structure of these communities can be used to differentiate objects handled by different individuals, even if those objects have been left untouched for up to 2 weeks at room temperature. Furthermore, we demonstrate that we can use a high-throughput pyrosequencing-based ap-proach to quantitatively compare the bacterial communities on objects and skin to match the object to the individual with a high degree of certainty. Although additional work is needed to further establish the utility of this approach, this series of studies introduces a forensics approach that could eventually be used to independently evaluate results obtained using more traditional forensic practices.	ENVO:human-associated habitat	TCAG	NA	UBERON:sebum	0	FWD:GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCCTGGCTCAG;REV:TGCTGCCTCCCGTAGGAGT	NA	private	20231444	y	Chris Lauber	CCME	GAZ:United States of America	human skin metagenome	0	NA	539655	n	NA	Fierer_forensic_keyboard	pyrosequencing	Forensic_identification_using_skin_bacterial_communities	FFCKVMW	ENVO:human-associated habitat	middle	FLX	Forensic_identification_using_skin_bacterial_communities	Forensic_identification_using_skin_bacterial_communities	fierer_forensic_keyboard	Forensic_identification_using_skin_bacterial_communities	CCME	male	years	232	16S_rRNA_gene_sequences_were_processed_according_to_the_methods_described_in_our_previous_publications_(Fierer_et_al.,_2008;_Hamady_et_al.,_2008)._Briefly,_sequences_<200_or_>300_nt_or_with_average_quality_scores_of_<25_were_removed_from_the_dataset,_as_were_those_with_uncorrectable_barcodes,_ambiguous_bases,_or_if_the_bacterial_16S_rRNA_gene-specific_primer_was_absent._Sequences_were_then_assigned_to_the_specific_subsamples_based_on_their_unique_12_nt_barcode_and_then_grouped_into_phylotypes_at_the_97%_level_of_sequence_identity_using_cd-hit_(Li_&_Godzik,_2006)_with_a_minimum_coverage_of_97%._We_chose_to_group_the_phylotypes_at_97%_identity_because_this_matches_the_limits_of_resolution_of_pyrosequencing_(Kunin_et_al.,_2010)_and_because_the_branch_length_so_omitted_contributes_little_to_the_tree_and_therefore_to_phylogenetic_estimates_of___diversity_(Hamady_et_al.,_2009)._A_representative_for_each_phylotype_was_chosen_by_selecting_the_most_abundant_sequence_in_the_phylotype,_with_ties_being_broken_by_choosing_the_longest_sequence._A_phylogenetic_tree_of_the_representative_sequences_was_constructed_using_the_Kimura_2-parameter_model_in_Fast_Tree_(Price_et_al.,_2009)_after_sequences_were_aligned_with_NAST_(minimum_150_nt_at_75%_minimum_identity)_(DeSantis_et_al.,_2006a)_against_the_GreenGenes_database_(DeSantis_et_al.,_2006b)._Hypervariable_regions_were_screened_out_of_the_alignment_using_PH_Lane_mask_(http://greengenes.lbl.gov/)._Differences_in_the_community_composition_for_each_pair_of_samples_were_determined_from_the_phylogenetic_tree_using_the_weighted_and_unweighted_UniFrac_algorithms_(Lozupone_&_Knight,_2005;_Lozupone_et_al.,_2006)._UniFrac_is_a_tree-based_metric_that_measures_the_distance_between_two_communities_as_the_fraction_of_branch_length_in_a_phylogenetic_tree_that_is_unique_to_one_of_the_communities_(as_opposed_to_being_shared_by_both)._This_method_of_community_comparison_accounts_for_the_relative_similarities_and_differences_among_phylotypes_(or_higher_taxa)_rather_than_treating_all_taxa_at_a_given_level_of_divergence_as_equal_(Lozupone_&_Knight,_2008)._Although_UniFrac_depends_on_a_phylogenetic_tree,_it_is_relatively_robust_to_differences_in_the_tree_reconstruction_method_or_to_the_approximation_of_using_phylotypes_to_represent_groups_of_very_similar_sequences_(Hamady_et_al.,_2009).	fierer_forensic_keyboard	NA	y	ENVO:sebum	16S rRNA	CCME	V2	36	NA	40.0083	Forensic_identification_using_skin_bacterial_communities	right middle finger tip