{"id":"W3213960624","doi":"10.33137/utjph.v2i2.36809","title":"Statistical Modelling of Population-Level Exonic Variant Frequency Data with an Emphasis on Rare Variants","year":2021,"lang":"en","type":"article","venue":"University of Toronto Journal of Public Health","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Lunenfeld-Tanenbaum Research Institute; Public Health Ontario; University of Toronto","funders":"","keywords":"Allele frequency; Population; Annotation; Genetics; Minor allele frequency; Biology; Allele; 1000 Genomes Project; Computational biology; Genotype; Medicine; Single-nucleotide polymorphism; Gene","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001030762,0.00007443529,0.0002815682,0.00002844692,0.00008541248,0.000005720736,0.0002955,0.00008792158,0.000131375],"category_scores_gemma":[0.000217812,0.00007245586,0.00003866225,0.00003886739,0.00004103876,0.00003154155,0.00007082089,0.00007596432,2.220708e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001147689,"about_ca_system_score_gemma":0.001170316,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006923377,"about_ca_topic_score_gemma":0.006515244,"domain_scores_codex":[0.9987358,0.0004027557,0.0003201297,0.0001872414,0.0001699952,0.0001840588],"domain_scores_gemma":[0.9984343,0.00004792762,0.0005341907,0.0003670559,0.0004231718,0.0001933755],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002396107,0.01008447,0.5126714,0.0009203539,0.006205875,0.0007885652,0.009633366,0.07468177,0.006830969,0.05355914,0.03015973,0.2920683],"study_design_scores_gemma":[0.004703515,0.006792421,0.9485546,0.0001848399,0.0001809336,0.0003580954,0.008342194,0.0192084,0.00005213272,0.0009138152,0.01022923,0.0004797617],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3940388,0.0007556644,0.6028147,0.001430683,0.00007928743,0.00006036027,0.0004943912,0.000001768692,0.0003244259],"genre_scores_gemma":[0.8686443,0.0007512373,0.1300029,0.00008165547,0.00004836029,2.336508e-8,0.0004140043,0.000005484479,0.00005208032],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.4746055,"threshold_uncertainty_score":0.9996896,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1168711424311615,"score_gpt":0.3096135271212686,"score_spread":0.1927423846901072,"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."}}