{"id":"W2015616366","doi":"10.1093/nar/gki394","title":"OrfPredictor: predicting protein-coding regions in EST-derived sequences","year":2005,"lang":"en","type":"article","venue":"Nucleic Acids Research","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":417,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University","funders":"Genome Canada","keywords":"Biology; Open reading frame; Coding region; Annotation; Coding (social sciences); Computational biology; Reading frame; Genetics; Peptide sequence; Gene; Mathematics","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":[0.001160934,0.002540491,0.001802246,0.002663183,0.0006970855,0.001319001,0.001463148,0.0009279576,0.01615312],"category_scores_gemma":[0.00169874,0.0011666,0.001167147,0.003053076,0.000282658,0.001190564,0.0007444325,0.001343786,0.02179948],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005800237,"about_ca_system_score_gemma":0.0009652026,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006714231,"about_ca_topic_score_gemma":0.001004095,"domain_scores_codex":[0.9995512,0.00006893736,0.00005658164,0.0001279959,0.0001384079,0.00005670849],"domain_scores_gemma":[0.9993653,0.0003060733,0.00008903851,0.00005437963,0.000110775,0.00007447019],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.005105824,0.0007058348,0.01293036,0.005860492,0.0003213401,0.003037613,0.0004369572,0.01201419,0.3651722,0.005489566,0.4009818,0.1879439],"study_design_scores_gemma":[0.0009583924,0.0007536041,0.02822588,0.0005260477,0.0004547114,0.004195741,0.0004299294,0.2094452,0.4179447,0.01067084,0.3260224,0.0003726049],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"methods","genre_scores_codex":[0.09025135,0.002255504,0.2407889,0.000466183,0.0004483916,0.0006029216,0.500417,0.154761,0.01000863],"genre_scores_gemma":[0.06295465,0.001206227,0.2454287,0.0001608226,0.0001295129,0.0004852661,0.6793133,0.005385933,0.004935451],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.01615312,"threshold_uncertainty_score":0.05403757,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04833340149248012,"score_gpt":0.3249245689733773,"score_spread":0.2765911674808972,"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."}}