{"id":"W4362562485","doi":"10.1093/nargab/lqad031","title":"Known sequence features explain half of all human gene ends","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Occupational Cancer Research Centre; University of Toronto","funders":"Canadian Institutes of Health Research; University of Toronto","keywords":"Polyadenylation; Gene; Biology; Genetics; Coding region; Computational biology; Sequence (biology); Primary transcript; Context (archaeology); RNA; Messenger RNA; Alternative splicing","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":[0.0004982805,0.0006908778,0.0005179509,0.0008042292,0.000339174,0.0005418907,0.0002859298,0.0006894121,0.003590745],"category_scores_gemma":[0.001442394,0.000225928,0.0009075967,0.0006196251,0.0002784384,0.0004355085,0.0003584013,0.0004353682,0.00182117],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002369931,"about_ca_system_score_gemma":0.0002916941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001721625,"about_ca_topic_score_gemma":0.002812275,"domain_scores_codex":[0.9996343,0.00009105226,0.00001726879,0.0001548593,0.00005417467,0.00004826737],"domain_scores_gemma":[0.9992365,0.0005084538,0.000057568,0.0001048418,0.00005625042,0.00003654369],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002392908,0.0002740393,0.3730685,0.0008240253,0.0008913514,0.002315869,0.0004182681,0.1899446,0.0981767,0.00564182,0.008859016,0.3171931],"study_design_scores_gemma":[0.00006844312,0.0003380558,0.2055072,0.0001370888,0.0004856175,0.00388499,0.0002366909,0.7234491,0.03221598,0.02026748,0.0133434,0.0000659798],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9087055,0.002819222,0.07639153,0.0003151555,0.0000503105,0.00004626872,0.005516494,0.0009914769,0.005164179],"genre_scores_gemma":[0.9899412,0.0003650462,0.004236158,0.00007107676,0.0000205086,0.00001942801,0.004053901,0.00008543221,0.001207234],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003590745,"threshold_uncertainty_score":0.01201218,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02065026306047327,"score_gpt":0.2755685505550601,"score_spread":0.2549182874945868,"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."}}