{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.06544594,0.002764611,0.002263831,0.01208231,0.003522974,0.007691872,0.004062884,0.003528057,0.01009847],"category_scores_gemma":[0.1270188,0.002008867,0.00260811,0.01059272,0.00366016,0.00981439,0.01238499,0.00701912,0.01247155],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00499786,"about_ca_system_score_gemma":0.01607962,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008135876,"about_ca_topic_score_gemma":0.007988164,"domain_scores_codex":[0.9642031,0.01130248,0.009719279,0.004716507,0.007950606,0.002108053],"domain_scores_gemma":[0.8459331,0.04888119,0.01857442,0.05646379,0.0259402,0.004207415],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.007904716,0.001252034,0.04705183,0.01008637,0.0005594034,0.001811669,0.01104005,0.008406655,0.08270133,0.116571,0.4274693,0.2851457],"study_design_scores_gemma":[0.0002921635,0.0002964713,0.01277061,0.003902422,0.0002206483,0.0005771727,0.001731808,0.006457094,0.06819549,0.03597199,0.8690065,0.0005777455],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0237323,0.001552501,0.6540385,0.005234024,0.002987748,0.005302942,0.1997509,0.08362654,0.02377458],"genre_scores_gemma":[0.04773435,0.001334837,0.6186602,0.003765185,0.0005184207,0.008209357,0.2943808,0.01707167,0.008325219],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9959371,"threshold_uncertainty_score":0.3461154,"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."}}