{"id":"W3009357978","doi":"10.1038/s41597-019-0320-2","title":"A reference library for Canadian invertebrates with 1.5 million barcodes, voucher specimens, and DNA samples","year":2019,"lang":"en","type":"article","venue":"Scientific Data","topic":"Environmental DNA in Biodiversity Studies","field":"Environmental Science","cited_by":110,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University; Public Health Agency of Canada; University of Guelph","funders":"Nature Conservancy of Canada; Canada First Research Excellence Fund; Ontario Genomics; Ontario Ministry of Research, Innovation and Science; Government of Canada; Churchill Northern Studies Centre; University of Guelph; Ministry of Environment; Canada Foundation for Innovation; Parks Canada; Genome Canada; Gordon and Betty Moore Foundation; Natural Sciences and Engineering Research Council of Canada; Smithsonian Institution","keywords":"GenBank; Biodiversity; Barcode; Biology; DNA barcoding; Taxonomic rank; Global biodiversity; Taxonomy (biology); Environmental DNA; DNA sequencing; Zoology; Ecology; Geography; Taxon; DNA; Genetics; Computer science","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002433097,0.001504825,0.001483939,0.01946969,0.00617836,0.002459786,0.004215548,0.001440652,0.03856495],"category_scores_gemma":[0.005368536,0.0008908445,0.000997924,0.02135837,0.0009849215,0.001022425,0.002094553,0.001577244,0.01825126],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01004826,"about_ca_system_score_gemma":0.04425985,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.663842,"about_ca_topic_score_gemma":0.8443224,"domain_scores_codex":[0.9969146,0.0001501747,0.0001979535,0.0005330262,0.001845808,0.000358296],"domain_scores_gemma":[0.9932148,0.0003024346,0.0004469984,0.0008363758,0.00467071,0.0005286897],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0006398771,0.0002807704,0.01167523,0.003527732,0.0002136218,0.0009087694,0.001455583,0.001350012,0.1175832,0.01211544,0.3702431,0.4800068],"study_design_scores_gemma":[0.00004992848,0.00007265154,0.02566871,0.0005698416,0.0001664037,0.0005147682,0.000260611,0.0007453895,0.01233076,0.001041608,0.9584876,0.00009163289],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.02880925,0.006973092,0.1388272,0.00162078,0.0009989478,0.00263458,0.7032376,0.01155258,0.1053459],"genre_scores_gemma":[0.02355921,0.003719695,0.2260142,0.0008458187,0.0001407905,0.001317344,0.6918564,0.001498213,0.05104822],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.336158,"threshold_uncertainty_score":0.6762754,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03989836595563699,"score_gpt":0.2084597603429137,"score_spread":0.1685613943872767,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}