{"id":"W4399538170","doi":"10.1099/mgen.0.001260","title":"PHA4GE quality control contextual data tags: standardized annotations for sharing public health sequence datasets with known quality issues to facilitate testing and training","year":2024,"lang":"en","type":"article","venue":"Microbial Genomics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University; Public Health Agency of Canada; Western University; Dalhousie University; Simon Fraser University","funders":"European and Developing Countries Clinical Trials Partnership; Biotechnology and Biological Sciences Research Council; U.S. National Library of Medicine; Foreign, Commonwealth and Development Office; Bill and Melinda Gates Foundation; European Commission; Medical Research Council; Directorate for Biological Sciences; Public Health Agency; National Institutes of Health; Public Health Agency of Canada","keywords":"Quality (philosophy); Computer science; Control (management); Sequence (biology); Data quality; Information retrieval; Data science; Knowledge management; Artificial intelligence; Engineering; Biology; Operations management","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":[0.00234294,0.0002032075,0.000361298,0.00005158378,0.0002521898,0.0002451902,0.0004096599,0.0001031601,0.000003220128],"category_scores_gemma":[0.001668382,0.0001797308,0.00003578035,0.0001304758,0.0002457337,0.0000192947,0.0002862424,0.000130164,0.000002243875],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005805307,"about_ca_system_score_gemma":0.000584138,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004364947,"about_ca_topic_score_gemma":0.0008910106,"domain_scores_codex":[0.9980191,0.0001431146,0.0004630621,0.0008178977,0.00009955477,0.0004572611],"domain_scores_gemma":[0.9987623,0.0002732315,0.0001022399,0.0005454418,0.0001084276,0.0002083377],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004091646,0.00004738651,0.0003524921,0.0003561363,0.000258679,0.00000508923,0.003919773,0.00003619542,0.68634,0.0002191684,0.00814353,0.2999124],"study_design_scores_gemma":[0.002470698,0.0008577466,0.0005712373,0.0001467252,0.00004128105,0.00004465772,0.002818122,0.001154646,0.00232128,0.0001047061,0.9888947,0.000574175],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7237449,0.004022912,0.1979034,0.006199332,0.0002374629,0.0009062304,0.06682698,0.0001266913,0.00003202611],"genre_scores_gemma":[0.8999485,0.0001069318,0.08986074,0.001690281,0.0003145156,0.0000534713,0.007794011,0.00004106019,0.0001905188],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9807512,"threshold_uncertainty_score":0.7329211,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3385397437314693,"score_gpt":0.4206014580559228,"score_spread":0.08206171432445347,"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."}}