{"id":"W4410833132","doi":"10.1101/2025.05.27.656002","title":"The Genetic and Evolutionary Landscape of Pentanucleotide Tandem Repeats in Human Genomes","year":2025,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Genetics and Neurodevelopmental Disorders","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Montreal Neurological Institute and Hospital","funders":"","keywords":"Tandem repeat; Genome; Evolutionary biology; Biology; Tandem; Computational biology; Genetics; Gene; Engineering","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":[0.0003833075,0.0001091722,0.0002138623,0.001306798,0.0002765985,0.0004573371,0.0001709655,0.0002624369,0.0006143835],"category_scores_gemma":[0.001553916,0.0001288201,0.0001639924,0.001294922,0.0002028981,0.0002326391,0.0002569812,0.0002118934,0.0001021105],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002557147,"about_ca_system_score_gemma":0.0001501322,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001885356,"about_ca_topic_score_gemma":0.002974955,"domain_scores_codex":[0.999627,0.0001067479,0.00001995463,0.0001591918,0.00005351719,0.00003358728],"domain_scores_gemma":[0.9994374,0.0002805735,0.0001486271,0.0000356818,0.00005593661,0.00004190917],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000449709,0.00006943294,0.8463769,0.0001189177,0.0003085177,0.0004565649,0.0009278952,0.008567314,0.09744631,0.002054853,0.000583947,0.04263953],"study_design_scores_gemma":[0.00001381415,0.00007565683,0.9525945,0.0000261863,0.00006671701,0.0008150594,0.0005333596,0.03659651,0.004466708,0.002829254,0.001957545,0.00002469234],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.997435,0.0002580934,0.001702693,0.00004237371,0.000001093632,0.000005669102,0.0002919048,0.000014446,0.0002487683],"genre_scores_gemma":[0.9958513,0.0001072385,0.003260843,0.00003235893,0.000004394919,0.00001024345,0.000617814,0.000009366076,0.0001065159],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001885356,"threshold_uncertainty_score":0.003748775,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00742958343180816,"score_gpt":0.2116278662051736,"score_spread":0.2041982827733654,"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."}}