{"id":"W2109550992","doi":"10.1093/bioinformatics/bth946","title":"Into the heart of darkness: large-scale clustering of human non-coding DNA","year":2004,"lang":"en","type":"article","venue":"Bioinformatics","topic":"Genome Rearrangement Algorithms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":69,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Human genome; Biology; Genome; Computational biology; Conserved sequence; Cluster analysis; Genetics; Gene; Coding region; Computer science; Peptide sequence; Artificial intelligence","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":[0.001017645,0.0003669184,0.0003785857,0.001662809,0.0007406157,0.0009195721,0.0007892904,0.0004772314,0.001488942],"category_scores_gemma":[0.005081486,0.0002186113,0.0002732179,0.001663785,0.0007844294,0.0007982517,0.0010697,0.0005954581,0.0004732619],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004733347,"about_ca_system_score_gemma":0.0004099766,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002181118,"about_ca_topic_score_gemma":0.00342855,"domain_scores_codex":[0.9996033,0.0001036572,0.00001505429,0.000140676,0.0001019462,0.0000352875],"domain_scores_gemma":[0.9984713,0.0008313038,0.0001660264,0.0002181046,0.0001662738,0.0001469295],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001426131,0.0003010341,0.09483663,0.001326978,0.0004474193,0.0008540548,0.002841427,0.1658921,0.1425827,0.07477777,0.0409715,0.4737422],"study_design_scores_gemma":[0.0001457887,0.0001637765,0.0795738,0.0001406187,0.000135552,0.0008149602,0.001326395,0.6188065,0.04175824,0.2320768,0.02495858,0.00009899122],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5475115,0.002139132,0.4404818,0.002098986,0.0001346393,0.0001349339,0.001843393,0.002491508,0.003164123],"genre_scores_gemma":[0.8110135,0.0005193157,0.1793771,0.0003558407,0.00009045187,0.00007028314,0.005651359,0.0003649044,0.002557229],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.002181118,"threshold_uncertainty_score":0.005381882,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009885294791323542,"score_gpt":0.2488597369418888,"score_spread":0.2389744421505653,"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."}}