{"id":"W4399390779","doi":"10.1038/s41597-024-03443-5","title":"The Pan-Canadian Chemical Library: A Mechanism to Open Academic Chemistry to High-Throughput Virtual Screening","year":2024,"lang":"en","type":"article","venue":"Scientific Data","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Winnipeg; University of Alberta; Structural Genomics Consortium; University of Toronto","funders":"National Institute of General Medical Sciences; Genentech; European Federation of Pharmaceutical Industries and Associations; University of Toronto; Ontario Genomics Institute; Merck KGaA; Alliance de recherche numérique du Canada; National Institutes of Health; Ontario Genomics; Genome Canada; McGill University; Bayer; Pfizer; Bristol-Myers Squibb","keywords":"Chemical space; Drug discovery; Workflow; Virtual screening; Computer science; Chemical library; Data science; Cheminformatics; Chemical database; Process (computing); Mechanism (biology); Chemistry; Chemical biology; Resource (disambiguation); Nanotechnology; Small molecule; Database; Physics; Materials science; Computational chemistry; Organic chemistry","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"methods","about_ca_system":true,"about_ca_topic":false,"confidence":"medium","status":"direct model label, unvalidated"},{"model":"gpt","categories":[],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":true,"about_ca_topic":true,"confidence":"low","status":"direct model label, unvalidated"}],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.001995221,0.001043509,0.0009690819,0.003121007,0.001822286,0.002083719,0.00296503,0.00084665,0.01122046],"category_scores_gemma":[0.004194284,0.0006553793,0.001065676,0.004597292,0.00155343,0.001477592,0.003858282,0.001910203,0.004812928],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003262876,"about_ca_system_score_gemma":0.01467369,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.09668458,"about_ca_topic_score_gemma":0.2084603,"domain_scores_codex":[0.9981862,0.0002355297,0.00006848502,0.0002476318,0.001040414,0.0002218022],"domain_scores_gemma":[0.9983789,0.0002402283,0.00008636755,0.0004284044,0.0005594063,0.0003067472],"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":[0.001817365,0.0008281171,0.003841647,0.001311022,0.0003214621,0.0008059817,0.0006500719,0.02262324,0.1328284,0.113314,0.2230472,0.4986115],"study_design_scores_gemma":[0.0005718056,0.0005198029,0.002720751,0.0001273599,0.0001769801,0.000600948,0.0001547486,0.0404607,0.1022909,0.02007047,0.8320172,0.0002883824],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05977739,0.004626623,0.6669582,0.00563918,0.0009583761,0.006851024,0.04047036,0.06443249,0.1502864],"genre_scores_gemma":[0.135583,0.004342275,0.7690748,0.002571586,0.0001288757,0.00295126,0.03862142,0.003232941,0.0434938],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.997035,"threshold_uncertainty_score":0.1922436,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06922418333834772,"score_gpt":0.3512207175254694,"score_spread":0.2819965341871217,"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."}}