{"id":"W4404976894","doi":"10.1186/s12859-024-05952-x","title":"Managing false positives during detection of pathogen sequences in shotgun metagenomics datasets","year":2024,"lang":"en","type":"article","venue":"BMC Bioinformatics","topic":"Salmonella and Campylobacter epidemiology","field":"Agricultural and Biological Sciences","cited_by":18,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Food Inspection Agency; Carleton University","funders":"Alliance de recherche numérique du Canada; Ontario Ministry of Agriculture, Food and Rural Affairs; Canadian Food Inspection Agency","keywords":"False positive paradox; Metagenomics; Shotgun; Computational biology; Shotgun sequencing; DNA microarray; Biology; True positive rate; Computer science; Bioinformatics; Genetics; DNA sequencing; Gene; Artificial intelligence; Gene expression","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":[],"consensus_categories":[],"category_scores_codex":[0.0003823029,0.0001083361,0.0001831179,0.00004192161,0.00005175127,0.00004305508,0.0001642468,0.00007714211,0.00002363212],"category_scores_gemma":[0.00003386366,0.0000455994,0.00007049718,0.0002920574,0.00004942106,0.0002762694,0.00007468729,0.00009116685,0.00002465699],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002898101,"about_ca_system_score_gemma":0.000007897713,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008046547,"about_ca_topic_score_gemma":0.0003807808,"domain_scores_codex":[0.9991059,0.00006309163,0.0004129096,0.0001308837,0.00008645236,0.0002008212],"domain_scores_gemma":[0.9995916,0.0002135964,0.00008552001,0.0000572746,0.00001242477,0.00003958082],"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":[0.00006799736,0.000046546,0.03541704,0.000591495,0.00005567692,0.00002677711,0.001273923,0.0004734202,0.8387955,0.000566198,0.00004004697,0.1226454],"study_design_scores_gemma":[0.0002838436,0.0003177288,0.7944415,0.0003645797,0.00008915659,0.0001660241,0.006458648,0.1393742,0.05147955,0.003426274,0.002892625,0.0007059354],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9979067,0.0006903392,0.0004748617,0.00006056362,0.0001668401,0.0001265584,0.0002872436,0.00004907457,0.0002378253],"genre_scores_gemma":[0.9977463,0.000250885,0.001676767,0.00004492907,0.00005614153,0.000006500167,0.0002014337,8.645775e-7,0.00001616017],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.787316,"threshold_uncertainty_score":0.185949,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02605075273000819,"score_gpt":0.2442765274455539,"score_spread":0.2182257747155457,"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."}}