{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.1186564,0.001696053,0.002545664,0.005419238,0.001234045,0.005895586,0.00428111,0.004403427,0.001191395],"category_scores_gemma":[0.4175566,0.00117939,0.002719501,0.004142166,0.006337654,0.004437109,0.005669403,0.004192731,0.0006708839],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002088725,"about_ca_system_score_gemma":0.002330312,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001328601,"about_ca_topic_score_gemma":0.0006963421,"domain_scores_codex":[0.9177266,0.05147313,0.00612169,0.01203493,0.01159784,0.001045901],"domain_scores_gemma":[0.4307628,0.4666737,0.03353721,0.05288783,0.01359499,0.002543475],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002791681,0.0004423082,0.2260858,0.001539255,0.00455469,0.001088931,0.001477119,0.4227032,0.00961624,0.08382526,0.008446319,0.2374292],"study_design_scores_gemma":[0.0001693769,0.0004442263,0.0217797,0.0002686057,0.0003908319,0.0006276454,0.000158335,0.7925504,0.01019605,0.1700776,0.003195768,0.0001414354],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.07094177,0.001820326,0.9228771,0.001480588,0.0001285903,0.0001500972,0.0006500112,0.000913935,0.00103752],"genre_scores_gemma":[0.6842171,0.0006398893,0.3105126,0.000734247,0.0005358591,0.000436931,0.001763672,0.0004611315,0.0006985551],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8813437,"threshold_uncertainty_score":0.6275222,"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."}}