{"id":"W2184129002","doi":"10.1186/s40793-015-0101-2","title":"Annotation inconsistencies beyond sequence similarity-based function prediction – phylogeny and genome structure","year":2015,"lang":"en","type":"article","venue":"Standards in Genomic Sciences","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":46,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"University of California, San Francisco; University of Toronto; University of Miami","keywords":"Annotation; Genome; Computational biology; Function (biology); Biology; Phylogenetic tree; Protein function; Similarity (geometry); Phylogenetics; Sequence (biology); Genome project; Evolutionary biology; Genetics; Data science; Computer science; Gene; Artificial intelligence","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04165426,0.001174677,0.001810507,0.005793508,0.002998708,0.01082006,0.005493928,0.003926098,0.002304395],"category_scores_gemma":[0.1338584,0.000928779,0.001113373,0.008971864,0.007798539,0.01687721,0.008012619,0.004185951,0.001568533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002486599,"about_ca_system_score_gemma":0.004452489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002317857,"about_ca_topic_score_gemma":0.001512032,"domain_scores_codex":[0.9445548,0.01639765,0.007785044,0.008286595,0.02189774,0.001078216],"domain_scores_gemma":[0.8162569,0.08141309,0.01823795,0.05626344,0.02666242,0.001166177],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0005687647,0.0001373435,0.0353695,0.001859395,0.0003310694,0.00133008,0.01053549,0.01141274,0.01758198,0.5726871,0.01205425,0.3361322],"study_design_scores_gemma":[0.00004065699,0.00009980948,0.009023177,0.001563933,0.0002350432,0.002150939,0.003021806,0.03007061,0.03896816,0.7898006,0.1247992,0.0002262032],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04788015,0.006409891,0.9159956,0.01412421,0.001017272,0.0001277762,0.001257507,0.003059766,0.01012782],"genre_scores_gemma":[0.4791865,0.006343016,0.4954691,0.006808354,0.001057445,0.0002661351,0.003905796,0.001935559,0.005028078],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.04165426,"threshold_uncertainty_score":0.2202914,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02904021147783195,"score_gpt":0.2779713684058358,"score_spread":0.2489311569280039,"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."}}