{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008414882,0.0001489907,0.0001399654,0.0001047685,0.000205574,0.00006287189,0.0001517311,0.00009814706,0.000004772262],"category_scores_gemma":[0.000127313,0.0001365619,0.00002817083,0.0001884491,0.0005074182,0.000004856855,0.00009458684,0.00006896401,4.309812e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000868612,"about_ca_system_score_gemma":0.0005676594,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003939955,"about_ca_topic_score_gemma":0.000319551,"domain_scores_codex":[0.9987627,0.00006837199,0.000213783,0.0004440131,0.0002806022,0.00023048],"domain_scores_gemma":[0.9994746,0.00001488152,0.00008530473,0.000159932,0.0001871656,0.0000780659],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","study_design_scores_codex":[0.0001137228,0.00001310182,0.05290711,0.00001380796,0.00001888547,0.00000128559,0.0004236127,0.007444873,0.9371698,0.0001770273,0.0001111932,0.001605536],"study_design_scores_gemma":[0.004585102,0.004849582,0.8235291,0.00004735041,0.0001174666,0.00006665924,0.007014635,0.004913803,0.05765378,0.03015695,0.06568664,0.001378901],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9944807,0.002820644,0.0006391917,0.0001224034,0.0003349115,0.0001887629,0.0006621139,0.000005489821,0.0007457867],"genre_scores_gemma":[0.9971527,0.0001223992,0.002268427,0.0002434221,0.0001290502,0.0000111229,0.00004375964,0.000008273969,0.00002086857],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8795161,"threshold_uncertainty_score":0.5568833,"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."}}