{"id":"W3027519114","doi":"10.1139/gen-2019-0206","title":"coil: an R package for cytochrome <i>c</i> oxidase I (COI) DNA barcode data cleaning, translation, and error evaluation","year":2020,"lang":"en","type":"article","venue":"Genome","topic":"Environmental DNA in Biodiversity Studies","field":"Environmental Science","cited_by":37,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Guelph","funders":"Government of Canada; Ontario Genomics; Compute Canada; Genome Canada","keywords":"Barcode; DNA barcoding; Biology; Indel; Computational biology; Genetics; Reading frame; Sequence analysis; Frameshift mutation; Error detection and correction; Frame (networking); GenBank; DNA; Computer science; Open reading frame; Gene; Algorithm; Evolutionary biology; Mutation; Peptide sequence","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003036923,0.0001486647,0.0001401125,0.00001282505,0.0002869247,0.00003333106,0.0003860779,0.00004846583,0.0007981212],"category_scores_gemma":[0.00004068828,0.0001556555,0.00002574792,0.0000897737,0.0002224364,0.0005313055,0.0003897484,0.00006366048,0.0002631489],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008654207,"about_ca_system_score_gemma":0.000003608448,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003445502,"about_ca_topic_score_gemma":0.00005252273,"domain_scores_codex":[0.9986246,0.00005845273,0.0001626828,0.0005935936,0.0003348897,0.0002258475],"domain_scores_gemma":[0.9993417,0.00003488006,0.00006338304,0.0003872413,0.000004969858,0.0001677866],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","study_design_scores_codex":[0.0001286249,0.0001561415,0.03140655,0.00006304179,0.00007016025,0.000007639278,0.009469989,0.0009731532,0.9379978,0.000009465784,0.004078812,0.01563862],"study_design_scores_gemma":[0.002578927,0.0006576455,0.9245796,0.000008209063,0.0002742002,0.000006130486,0.001794846,0.01883918,0.002903215,0.0002490405,0.04740291,0.000706086],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9935257,0.0002903931,0.001960777,0.001585658,0.00003255194,0.0009483402,0.0009362,0.00005853411,0.0006617822],"genre_scores_gemma":[0.9856163,0.00005605649,0.01248934,0.001065206,0.0000609616,0.00002376697,0.0006338367,0.00001798527,0.00003660424],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9350946,"threshold_uncertainty_score":0.8738867,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1242940691951622,"score_gpt":0.2875982603984256,"score_spread":0.1633041912032634,"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."}}