{"id":"W4389174422","doi":"10.1038/s41588-023-01524-6","title":"Benchmarking of deep neural networks for predicting personal gene expression from DNA sequence highlights shortcomings","year":2023,"lang":"en","type":"article","venue":"Nature Genetics","topic":"Genomics and Chromatin Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":114,"is_retracted":false,"has_abstract":false,"ca_institutions":"Canadian Institute for Advanced Research","funders":"National Institute on Aging","keywords":"Benchmarking; Biology; Computational biology; DNA sequencing; Gene; Genetics; Personal genomics; Genome; Genomics; Human genome","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002583861,0.001010824,0.0007570628,0.0005688228,0.0003381669,0.001311306,0.001507782,0.001122453,0.002574642],"category_scores_gemma":[0.007424337,0.0003632396,0.000541034,0.0009036383,0.0004456858,0.001495798,0.0008344766,0.001377579,0.000953212],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00118039,"about_ca_system_score_gemma":0.001367687,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01609948,"about_ca_topic_score_gemma":0.01837395,"domain_scores_codex":[0.9988691,0.0003586943,0.00007026607,0.0002827767,0.0002893517,0.0001297292],"domain_scores_gemma":[0.99765,0.001097964,0.00006286418,0.0004556282,0.0006366954,0.00009673777],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0004659564,0.0002191179,0.008397292,0.0001806171,0.0002724783,0.0000771071,0.00004464336,0.7536009,0.008153098,0.003583222,0.01014511,0.2148605],"study_design_scores_gemma":[0.00001187487,0.00005786545,0.0009407851,0.00001430494,0.00001595654,0.00001451526,0.00001677546,0.9913374,0.004425724,0.002286268,0.0008713358,0.000007081616],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7024115,0.005439427,0.2591761,0.003187693,0.0007971886,0.0001071975,0.003642075,0.01076913,0.01446971],"genre_scores_gemma":[0.9441494,0.0006193793,0.04654577,0.0003456342,0.00005891792,0.00006391645,0.004399131,0.0002618273,0.003556065],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01609948,"threshold_uncertainty_score":0.03201157,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.008019904422525752,"score_gpt":0.2400583485352147,"score_spread":0.232038444112689,"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."}}