{"id":"W4416421739","doi":"10.1093/bib/bbaf611","title":"Precision in prediction: tailoring machine learning models for breast cancer missense variants pathogenicity prediction","year":2025,"lang":"en","type":"article","venue":"Briefings in Bioinformatics","topic":"Genomics and Rare Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"United Arab Emirates University","keywords":"Interpretability; Benchmarking; Breast cancer; Pathogenicity; Precision medicine; Feature (linguistics); Robustness (evolution); Dimensionality reduction","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002709013,0.0001500385,0.000157803,0.0001366613,0.0001042048,0.00004188446,0.0001228782,0.0001632115,0.000004984418],"category_scores_gemma":[0.00009750744,0.0001518313,0.00007241149,0.000174397,0.00002612952,0.00002532876,0.0001067646,0.000125375,5.77387e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006281776,"about_ca_system_score_gemma":0.0001315045,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001609506,"about_ca_topic_score_gemma":0.0001027799,"domain_scores_codex":[0.998959,0.00002180151,0.0004643709,0.0002127062,0.00009728824,0.0002448088],"domain_scores_gemma":[0.9995537,0.00002132734,0.0001113928,0.0001750451,0.00008917899,0.00004930732],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002930157,0.0007778989,0.106036,0.001711837,0.0003007324,0.00002502558,0.00232389,0.6792307,0.05443804,0.001360414,0.005471178,0.1453941],"study_design_scores_gemma":[0.00217991,0.0001067812,0.04282767,0.000343367,0.00003714108,0.00004529452,0.0001941341,0.9465654,0.002001041,0.0007235918,0.004729768,0.0002459052],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9266722,0.001433105,0.06761044,0.000438819,0.0004745121,0.0009086793,0.001486137,0.0000437649,0.0009323033],"genre_scores_gemma":[0.9946457,0.00116931,0.0030775,0.0003642581,0.00008088905,0.00009384814,0.0003296226,0.00001644213,0.0002224503],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2673347,"threshold_uncertainty_score":0.6191502,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01136543101514288,"score_gpt":0.237825636749706,"score_spread":0.2264602057345631,"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."}}