{"id":"W840861454","doi":"10.1371/journal.pone.0129711","title":"The Discovery of Novel Biomarkers Improves Breast Cancer Intrinsic Subtype Prediction and Reconciles the Labels in the METABRIC Data Set","year":2015,"lang":"en","type":"article","venue":"PLoS ONE","topic":"Breast Cancer Treatment Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":36,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"BC Cancer Agency; Australian Research Council; Cancer Institute NSW; Cancer Research UK","keywords":"Breast cancer; Artificial intelligence; Classifier (UML); Computer science; Random forest; Machine learning; Ensemble learning; Computational biology; Discriminative model; Support vector machine; Transcriptome; Bioinformatics; Cancer; Medicine; Gene; Gene expression; Biology; Internal medicine","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":[0.003543761,0.0007632482,0.0007536214,0.002444382,0.0005489863,0.001168272,0.0006166759,0.0007093178,0.0008505227],"category_scores_gemma":[0.007546742,0.0001307393,0.0007410361,0.001372207,0.0002732127,0.0006486746,0.0009415218,0.001100727,0.0005258512],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005268644,"about_ca_system_score_gemma":0.0006860809,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001928452,"about_ca_topic_score_gemma":0.003207641,"domain_scores_codex":[0.998162,0.0004742191,0.0001389669,0.0006060185,0.0004362366,0.0001826167],"domain_scores_gemma":[0.9961616,0.002047706,0.0006306847,0.0004533009,0.0005588825,0.0001478112],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00118563,0.0003331964,0.6987916,0.0003914728,0.0005603534,0.0002365842,0.0003491759,0.03684266,0.03832266,0.0008162429,0.004654253,0.2175161],"study_design_scores_gemma":[0.00006429234,0.0006818096,0.5326319,0.0001232797,0.000558093,0.0007419464,0.000468709,0.4065469,0.04415724,0.004120649,0.009794677,0.0001105106],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9531413,0.00104528,0.03990339,0.0004029981,0.00003867505,0.00005932835,0.003294161,0.0006726609,0.001442249],"genre_scores_gemma":[0.9509909,0.0002195897,0.03890539,0.0001168953,0.00004627326,0.00008126764,0.009031862,0.00006935457,0.0005384714],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003543761,"threshold_uncertainty_score":0.01874143,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08054890599048548,"score_gpt":0.2708805444404884,"score_spread":0.1903316384500029,"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."}}