{"id":"W4285084440","doi":"10.1038/s41587-022-01369-0","title":"Standardized annotation of translated open reading frames","year":2022,"lang":"en","type":"letter","venue":"Nature Biotechnology","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":250,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Sherbrooke","funders":"National Cancer Institute; National Human Genome Research Institute; National Institute on Aging; RIKEN; Agència de Gestió d'Ajuts Universitaris i de Recerca; Medical Research Council; Canadian Institutes of Health Research; Directorate for Biological Sciences; National Institutes of Health; European Molecular Biology Laboratory; Russian Science Foundation; Institució Catalana de Recerca i Estudis Avançats; National Institute of General Medical Sciences; University College Cork; University of Leeds; Universitat Pompeu Fabra; Fondation Leducq; Stowers Institute for Medical Research; Australian Government; Searle Scholars Program; National Science Foundation; Science Foundation Ireland; Biotechnology and Biological Sciences Research Council; Université de Sherbrooke; Wellcome Trust; European Commission; University of California, Irvine; Broad Institute; University of Pittsburgh; Howard Hughes Medical Institute; Staatssekretariat für Bildung, Forschung und Innovation; Musella Foundation For Brain Tumor Research and Information; Agencia Estatal de Investigación; Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology; Alex's Lemonade Stand Foundation for Childhood Cancer; Yale University","keywords":"Annotation; Open reading frame; Reading (process); Computer science; Information retrieval; Computational biology; Natural language processing; Biology; Artificial intelligence; Linguistics; Genetics; Philosophy; Peptide sequence","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"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":[0.0071533,0.0007279761,0.0008087646,0.002858799,0.001162072,0.002156815,0.00156726,0.002574262,0.003296503],"category_scores_gemma":[0.01413657,0.0004299335,0.0005875914,0.002580398,0.001446949,0.00165101,0.001346365,0.005032898,0.009486219],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002477055,"about_ca_system_score_gemma":0.00318193,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00145654,"about_ca_topic_score_gemma":0.003699841,"domain_scores_codex":[0.9929289,0.002218407,0.001180352,0.000928584,0.002367863,0.0003758318],"domain_scores_gemma":[0.9837516,0.004778243,0.001850779,0.003373828,0.005642054,0.0006035283],"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":[0.0003485095,0.0001093698,0.003219853,0.001190983,0.00003970233,0.001974623,0.0005121077,0.0007346225,0.08339601,0.06544748,0.4837328,0.3592939],"study_design_scores_gemma":[0.00002464286,0.0001050502,0.002354124,0.0003061919,0.00002028027,0.001242511,0.0001709068,0.001626858,0.02074423,0.01449096,0.9588595,0.00005480846],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0322214,0.05416159,0.3994699,0.3290198,0.0579285,0.001455183,0.02582556,0.007636494,0.09228159],"genre_scores_gemma":[0.08831321,0.03407707,0.5939028,0.140617,0.02244526,0.001804017,0.06970236,0.00266763,0.04647072],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0071533,"threshold_uncertainty_score":0.03783071,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009975389999421257,"score_gpt":0.2940358937373858,"score_spread":0.2840605037379645,"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."}}