{"id":"W2938899814","doi":"10.1101/614560","title":"Creating reproducible pharmacogenomic analysis pipelines","year":2019,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; Institute of Cancer Research; Ontario Institute for Cancer Research; University Health Network; University of Toronto; Princess Margaret Cancer Centre","funders":"","keywords":"Workflow; Computer science; Scalability; Pipeline (software); Identifier; Data science; Pharmacogenomics; Pipeline transport; Process (computing); Field (mathematics); Data mining; Database; Bioinformatics; Engineering; Programming language","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":["metaepi_narrow","scholarly_communication","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.01521473,0.0006970399,0.001308309,0.002354361,0.0004373765,0.00274024,0.003910592,0.0003054909,0.0006479545],"category_scores_gemma":[0.005332406,0.0006323266,0.0007185714,0.005544392,0.0001458663,0.000360051,0.004288124,0.0007392898,0.002174444],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000251378,"about_ca_system_score_gemma":0.0005436147,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002030302,"about_ca_topic_score_gemma":0.000005096842,"domain_scores_codex":[0.9893903,0.0005091314,0.001887433,0.005007128,0.002360749,0.0008452724],"domain_scores_gemma":[0.9862455,0.0007561989,0.001558128,0.009604072,0.001472798,0.0003632731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001547349,0.0009527391,0.3705904,0.0005728714,0.005873094,0.0002353225,0.0001852015,0.1707006,0.3210499,0.001194429,0.1282127,0.0002780151],"study_design_scores_gemma":[0.001577649,0.00008279242,0.3486691,0.0004611851,0.004806021,3.233116e-8,0.00008433523,0.3381488,0.1062675,0.00007252581,0.1955489,0.004281241],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9430124,0.001299917,0.04456077,0.0005542444,0.008049647,0.0009985557,0.0005554685,0.0006577807,0.0003112504],"genre_scores_gemma":[0.9889603,0.00006648537,0.009111177,0.0002797117,0.0009586496,0.00005238293,0.000001589819,0.00007740674,0.0004923309],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2147824,"threshold_uncertainty_score":0.9996128,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0729836126310226,"score_gpt":0.330258002445866,"score_spread":0.2572743898148434,"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."}}