{"id":"W3112134181","doi":"10.1186/s13062-020-00287-y","title":"Systematic evaluation of supervised machine learning for sample origin prediction using metagenomic sequencing data","year":2020,"lang":"en","type":"article","venue":"Biology Direct","topic":"Microbial Community Ecology and Physiology","field":"Environmental Science","cited_by":14,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Agency of Canada","funders":"","keywords":"Metagenomics; Lasso (programming language); Artificial intelligence; Regression; Machine learning; Biology; Sample (material); Shotgun sequencing; Multivariate statistics; Sample size determination; Computer science; Statistics; Data mining; DNA sequencing; Mathematics; Genetics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02818976,0.002903335,0.001853902,0.002203846,0.0008742103,0.001441034,0.002567329,0.001701627,0.0007899867],"category_scores_gemma":[0.04278307,0.0006281218,0.001902145,0.001412257,0.001038838,0.001825719,0.001833486,0.00198507,0.0005637962],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00109816,"about_ca_system_score_gemma":0.002782064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003981561,"about_ca_topic_score_gemma":0.004042942,"domain_scores_codex":[0.9824125,0.01137893,0.001197087,0.002632781,0.002063032,0.0003156048],"domain_scores_gemma":[0.9447612,0.04059285,0.002431276,0.004225861,0.007140992,0.0008479519],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"systematic_review","study_design_scores_codex":[0.004277724,0.002713192,0.06755127,0.004163807,0.003384775,0.0004342553,0.0003510283,0.5328804,0.01328987,0.001447085,0.01060204,0.3589046],"study_design_scores_gemma":[0.0001444975,0.0006413401,0.004945931,0.0001409722,0.0002012094,0.00006752118,0.00007468063,0.9864091,0.005482044,0.0009516485,0.000907674,0.0000334431],"study_design_candidate":"systematic_review","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"review","genre_scores_codex":[0.7649457,0.01500482,0.202594,0.001166351,0.0005732537,0.001200699,0.004457805,0.007095293,0.002962081],"genre_scores_gemma":[0.8145649,0.001274566,0.1705527,0.0003663219,0.0001858846,0.0006452227,0.01140026,0.0003752068,0.0006349348],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.02818976,"threshold_uncertainty_score":0.1490835,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2225167887248018,"score_gpt":0.3360370779095436,"score_spread":0.1135202891847418,"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."}}