{"id":"W2096586490","doi":"10.1186/1756-0500-4-267","title":"GO Trimming: Systematically reducing redundancy in large Gene Ontology datasets","year":2011,"lang":"en","type":"article","venue":"BMC Research Notes","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":90,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Victoria","funders":"","keywords":"Trimming; Computer science; Redundancy (engineering); Gene ontology; Data mining; Information retrieval; Machine learning; Gene; Biology","routes":{"ca_aff":true,"ca_fund":false,"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":[0.008546658,0.001630737,0.001723591,0.00761417,0.00210349,0.001998677,0.002140796,0.000799055,0.002209673],"category_scores_gemma":[0.02431197,0.0006146609,0.002585087,0.006314888,0.001110315,0.001522371,0.003324998,0.002220104,0.001463805],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001002435,"about_ca_system_score_gemma":0.003343525,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002878625,"about_ca_topic_score_gemma":0.006374673,"domain_scores_codex":[0.9951259,0.001713077,0.0004861272,0.001022732,0.001363825,0.000288219],"domain_scores_gemma":[0.9901729,0.004437202,0.00110149,0.002547914,0.001445261,0.0002952092],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002122254,0.0008517155,0.05009969,0.005638331,0.002094625,0.001705159,0.003242382,0.02236943,0.2006696,0.0112299,0.0396786,0.6602983],"study_design_scores_gemma":[0.0009176022,0.002092229,0.1264868,0.001674361,0.003139968,0.00328477,0.003341564,0.4053434,0.2139608,0.1172999,0.121818,0.0006405709],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1929614,0.001923234,0.7725577,0.0009242306,0.0002821809,0.001206408,0.01067919,0.01649537,0.002970269],"genre_scores_gemma":[0.1470892,0.0006707046,0.8210686,0.0005063968,0.0001254791,0.001381137,0.02656613,0.001749493,0.0008429409],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008546658,"threshold_uncertainty_score":0.04519963,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1295681124709397,"score_gpt":0.369263634827619,"score_spread":0.2396955223566794,"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."}}