{"id":"W4281291437","doi":"10.1101/2022.05.06.22273576","title":"Datasets for a Simulated Family-Based Exome-Sequencing Study","year":2022,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Genetic factors in colorectal cancer","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada; Compute Canada","keywords":"Exome sequencing; Exome; Inheritance (genetic algorithm); Computational biology; Table (database); Population; Genetics; Scripting language; Biology; Computer science; Data mining; Mutation; Gene; Medicine; Programming language","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":[0.002732808,0.0005884938,0.0005476335,0.0009858406,0.0006844094,0.0009057631,0.001469269,0.001638145,0.02514753],"category_scores_gemma":[0.01053352,0.0004712231,0.001072533,0.002058944,0.0004109559,0.0004650346,0.0007490105,0.001367324,0.00348053],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009928548,"about_ca_system_score_gemma":0.0009117711,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009646741,"about_ca_topic_score_gemma":0.0123804,"domain_scores_codex":[0.9992704,0.0004041721,0.00005469766,0.0001244983,0.0001063793,0.00003989024],"domain_scores_gemma":[0.9933564,0.004851502,0.0001819238,0.0008339604,0.0005419102,0.0002343934],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002314306,0.0006058061,0.0629452,0.001019,0.0008156623,0.001677512,0.0003285134,0.6314024,0.002724081,0.01775368,0.2572611,0.02115281],"study_design_scores_gemma":[0.003628778,0.0005458267,0.06701783,0.0003523497,0.0003726157,0.001654703,0.000597958,0.6397013,0.005305064,0.07803961,0.202529,0.0002549138],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.172759,0.0006071807,0.03224922,0.002885364,0.0002513294,0.0006205901,0.7814068,0.002760184,0.006460201],"genre_scores_gemma":[0.2341443,0.0004238154,0.04483135,0.0006029856,0.00006063336,0.002127563,0.7131471,0.0005524346,0.00410989],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.02514753,"threshold_uncertainty_score":0.08412689,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09485746207486503,"score_gpt":0.3653223640327752,"score_spread":0.2704649019579102,"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."}}