{"id":"W3162765094","doi":"10.1101/2021.05.21.445109","title":"Analyzing Biomarker Discovery: Estimating the Reproducibility of Biomarker Sets","year":2021,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian VIGOUR Centre; University of Waterloo; University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Alberta Machine Intelligence Institute","keywords":"Reproducibility; Univariate; Biomarker; Biomarker discovery; Computer science; Binary number; False discovery rate; Statistics; Data mining; Pattern recognition (psychology); Artificial intelligence; Mathematics; Machine learning; Multivariate statistics; Biology; Proteomics","routes":{"ca_aff":true,"ca_fund":true,"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":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.02541959,0.0008487205,0.00216157,0.0002110275,0.0002498624,0.0004848316,0.001395443,0.0008309024,0.0001495228],"category_scores_gemma":[0.2639602,0.0006670405,0.0007434961,0.00122497,0.000736648,0.0002252483,0.00228811,0.00137725,0.000009677118],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002620305,"about_ca_system_score_gemma":0.0008293983,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000804208,"about_ca_topic_score_gemma":0.000002983186,"domain_scores_codex":[0.9885758,0.003323263,0.002977136,0.003397787,0.0009450529,0.0007810307],"domain_scores_gemma":[0.9633247,0.02239505,0.002554325,0.01019135,0.001247325,0.0002872821],"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":[0.0005604294,0.002811886,0.05426444,0.01421481,0.006163031,0.0003566125,0.0001413644,0.0001850518,0.8954192,0.02239413,0.003002473,0.0004865136],"study_design_scores_gemma":[0.002706232,0.0001706706,0.5110065,0.009406033,0.004490869,2.086805e-7,0.00005253769,0.02465005,0.4178009,0.02505294,0.0002744898,0.00438865],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7383804,0.0007163411,0.2539074,0.0005602606,0.003841564,0.001577253,0.0006687665,0.0003230128,0.00002506339],"genre_scores_gemma":[0.5262742,0.00003048534,0.4731492,0.00005087026,0.0002599256,0.0001128697,2.096754e-7,0.0001193331,0.000002898956],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4776184,"threshold_uncertainty_score":0.9995781,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2415899713479743,"score_gpt":0.435299415521316,"score_spread":0.1937094441733417,"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."}}