{"id":"W4396816249","doi":"10.48550/arxiv.2405.04734","title":"The Canadian VirusSeq Data Portal & Duotang: open resources for SARS-CoV-2 viral sequences and genomic epidemiology","year":2024,"lang":"en","type":"preprint","venue":"PubMed","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; AIDS Vancouver; Michael Smith Health Research BC; Genome Canada; University of Calgary; Sunnybrook Health Science Centre; BC Centre for Disease Control; Women and Children’s Health Research Institute; Sunnybrook Hospital; Ontario Genomics; University of Alberta; Indoc Research; Université de Montréal; McGill Genome Centre; University of Toronto; Western University; McGill University; Simon Fraser University; Public Health Agency of Canada; Ontario Institute for Cancer Research; University of British Columbia; Montreal Heart Institute","funders":"","keywords":"Interoperability; Pandemic; Public health; Data sharing; Data science; Genome; Genomics; Acknowledgement; World Wide Web; Coronavirus disease 2019 (COVID-19); Computer science; Biology; Medicine; Genetics","routes":{"ca_aff":true,"ca_fund":false,"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":["open_science"],"consensus_categories":[],"category_scores_codex":[0.005005824,0.001193432,0.0008512699,0.007433233,0.004528829,0.008192771,0.002951598,0.001577564,0.06850883],"category_scores_gemma":[0.02195521,0.0009866758,0.000860236,0.01401914,0.001719521,0.005085209,0.006904738,0.002196367,0.02650385],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01414009,"about_ca_system_score_gemma":0.05724498,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.8040327,"about_ca_topic_score_gemma":0.849249,"domain_scores_codex":[0.9960031,0.0004484412,0.0002644288,0.0004973873,0.002167534,0.0006190545],"domain_scores_gemma":[0.9805239,0.003366249,0.0007058714,0.004985119,0.006573296,0.00384554],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000270115,0.00004774929,0.003692458,0.0004296367,0.00002757283,0.0002358161,0.001332129,0.000812213,0.00154666,0.01557004,0.8837742,0.0922615],"study_design_scores_gemma":[0.00002254789,0.000006361381,0.002461888,0.0001088951,0.000005501129,0.0000337495,0.0003101768,0.0004178398,0.00049085,0.00220413,0.9938775,0.00006054229],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"software","genre_scores_codex":[0.006566013,0.001087682,0.03451592,0.007497146,0.000882416,0.001023309,0.7176302,0.0487583,0.1820389],"genre_scores_gemma":[0.04276442,0.002509389,0.1282267,0.003598234,0.0003491248,0.0008400358,0.7220877,0.01603075,0.08359361],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.9970484,"threshold_uncertainty_score":0.3942428,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1396706465663539,"score_gpt":0.3372018918401569,"score_spread":0.1975312452738029,"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."}}