{"id":"W4399850991","doi":"10.1101/2024.06.19.24309173","title":"A comprehensive tandem repeat catalog of the human genome","year":2024,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Genetic Neurodegenerative Diseases","field":"Neuroscience","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"BC Children's Hospital; University of British Columbia; Canada's Michael Smith Genome Sciences Centre","funders":"Canadian Institutes of Health Research; Alliance de recherche numérique du Canada","keywords":"Tandem repeat; Genome; Human genome; Genotyping; Biology; Microsatellite; Genetics; Computational biology; STR multiplex system; Gene; Genotype; Allele","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":[0.000954684,0.001057789,0.001135819,0.004904014,0.0009425859,0.001308426,0.0007064766,0.0006799398,0.008565637],"category_scores_gemma":[0.002194556,0.0003515696,0.0006595813,0.009803663,0.0002541082,0.000664121,0.001071252,0.0008521565,0.00908964],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003827944,"about_ca_system_score_gemma":0.002835365,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003212449,"about_ca_topic_score_gemma":0.005417848,"domain_scores_codex":[0.9992651,0.00008205677,0.00008240985,0.0002570201,0.000246949,0.00006638877],"domain_scores_gemma":[0.9988989,0.0002179352,0.0001877579,0.0002615664,0.0002688959,0.0001649823],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.001561854,0.0001888259,0.03636196,0.007798631,0.0007906477,0.001948977,0.001115157,0.003624661,0.3234274,0.01014392,0.2753706,0.3376673],"study_design_scores_gemma":[0.0001784806,0.0002171888,0.115595,0.0006694245,0.0006205515,0.003452918,0.000322672,0.003412637,0.04547842,0.008597996,0.8213027,0.0001520211],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.05454858,0.009248727,0.02693074,0.000447756,0.0001583037,0.0001667808,0.8965261,0.004030484,0.007942542],"genre_scores_gemma":[0.02577023,0.003216173,0.03431579,0.00013962,0.00007102114,0.0001659989,0.9336774,0.000476417,0.002167248],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.008565637,"threshold_uncertainty_score":0.02865487,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06007383145162026,"score_gpt":0.3001326257769286,"score_spread":0.2400587943253084,"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."}}