{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","open_science","research_integrity"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.0004819045,0.000355363,0.0007165098,0.0009059053,0.0001487832,0.0001589349,0.006454527,0.006282832,0.00007929479],"category_scores_gemma":[0.0002799554,0.0003285433,0.0001320371,0.001528502,0.000162343,0.0003695744,0.0014012,0.009540925,0.00000218782],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001787775,"about_ca_system_score_gemma":0.0002430428,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005279392,"about_ca_topic_score_gemma":0.000003095145,"domain_scores_codex":[0.9974654,0.0001746524,0.0004586306,0.0009186135,0.0005709313,0.0004118087],"domain_scores_gemma":[0.9977127,0.0001598771,0.0005746122,0.001341462,0.0001904793,0.00002083025],"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":[0.00004769458,0.00002601706,0.000002101182,0.0001976164,0.0001104974,0.0005175023,0.0002663506,6.589833e-7,0.01019492,0.04400048,0.8677922,0.07684401],"study_design_scores_gemma":[0.0002872981,0.0002038048,6.200337e-7,0.000103071,0.0000245942,0.00007469463,0.000006143201,0.0001195111,0.1159521,0.06953304,0.8133399,0.0003552197],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"methods","genre_scores_codex":[0.00005336692,0.0148349,0.1378234,0.8428898,0.0008698941,0.00083982,0.000153529,0.002129034,0.0004061913],"genre_scores_gemma":[0.003995528,0.0001642132,0.7777203,0.2168825,0.0002317001,0.0001016189,0.0003594149,0.00007674429,0.0004679052],"genre_candidate":"commentary","genre_consensus":null,"teacher_disagreement_score":0.639897,"threshold_uncertainty_score":0.9999167,"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."}}