{"id":"W2782879881","doi":"10.1099/mgen.0.000151","title":"Comprehensive assessment of the quality of Salmonella whole genome sequence data available in public sequence databases using the Salmonella in silico Typing Resource (SISTR)","year":2018,"lang":"en","type":"article","venue":"Microbial Genomics","topic":"Salmonella and Campylobacter epidemiology","field":"Agricultural and Biological Sciences","cited_by":63,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Agency of Canada","funders":"Public Health Agency; Public Health Agency of Canada","keywords":"In silico; Multilocus sequence typing; Salmonella; Genome; Whole genome sequencing; Biology; Typing; Serotype; Computational biology; Concordance; Genetics; Database; 1000 Genomes Project; Data quality; Data mining; Computer science; Gene; Microbiology; Genotype; Engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.01204037,0.0006355963,0.0009185756,0.005188711,0.0009513057,0.002172216,0.0006389703,0.00063927,0.00120725],"category_scores_gemma":[0.02558431,0.0005120804,0.0009974835,0.004476471,0.0006393273,0.001391866,0.001940017,0.0008175385,0.0008257283],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005491826,"about_ca_system_score_gemma":0.001740307,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003196805,"about_ca_topic_score_gemma":0.007013964,"domain_scores_codex":[0.9928182,0.001941399,0.001272106,0.001156753,0.002433989,0.000377589],"domain_scores_gemma":[0.9702803,0.009504649,0.006428488,0.004605807,0.008462458,0.0007181615],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001507197,0.0005596964,0.5373476,0.002608367,0.001341317,0.00064163,0.00287439,0.00679006,0.3249773,0.001046682,0.004151746,0.1161539],"study_design_scores_gemma":[0.00006319104,0.001141593,0.8283691,0.0007665702,0.0009004608,0.001208639,0.001746236,0.02021941,0.1156734,0.0009990279,0.02868624,0.0002261187],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9568256,0.001713223,0.0191477,0.0002035456,0.00004705226,0.0002027842,0.01931623,0.0006416998,0.001902078],"genre_scores_gemma":[0.8598877,0.001325324,0.05818654,0.0001849899,0.00003269996,0.0001682501,0.07914821,0.0003215745,0.0007447862],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01204037,"threshold_uncertainty_score":0.0636763,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2991486640787514,"score_gpt":0.3640092327917008,"score_spread":0.06486056871294937,"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."}}