{"id":"W4413615921","doi":"10.1101/2025.08.21.671542","title":"Systematic Evaluation of Somatic Contamination in Germline Genomes","year":2025,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"CRISPR and Genetic Engineering","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Calgary","funders":"National Natural Science Foundation of China","keywords":"Germline; Somatic cell; Contamination; Biology; Genome; Genetics; Germline mutation; Computational biology; Mutation; Ecology; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02756378,0.0006622589,0.0008503748,0.002365246,0.0009994607,0.002052805,0.001110449,0.0009889798,0.001484126],"category_scores_gemma":[0.07392693,0.000444649,0.001228361,0.002043544,0.001291471,0.0007600613,0.002032785,0.000918313,0.0003096531],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007817761,"about_ca_system_score_gemma":0.001649514,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001675541,"about_ca_topic_score_gemma":0.002908549,"domain_scores_codex":[0.9766154,0.01202243,0.002209447,0.004881798,0.003694616,0.0005762159],"domain_scores_gemma":[0.9063641,0.07044709,0.007788728,0.009751521,0.004754928,0.0008937219],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001957846,0.0002407996,0.5904276,0.002143851,0.005769945,0.00223932,0.001107195,0.03743789,0.1583688,0.02035581,0.004310183,0.1756408],"study_design_scores_gemma":[0.00037557,0.001674596,0.4476913,0.0007607873,0.004391413,0.004675012,0.0007106183,0.2414136,0.2214291,0.04596435,0.03064476,0.0002689374],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6623513,0.00265888,0.3259766,0.0005924543,0.0001661353,0.0003774117,0.003408893,0.001952486,0.002515764],"genre_scores_gemma":[0.9225978,0.000266866,0.07361489,0.0003441653,0.00003450021,0.0001278028,0.002479772,0.000283316,0.0002508883],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02756378,"threshold_uncertainty_score":0.1457729,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01073057746891761,"score_gpt":0.2745766710921844,"score_spread":0.2638460936232668,"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."}}