{"id":"W3120891503","doi":"10.1101/2021.01.04.425285","title":"debar, a sequence-by-sequence denoiser for COI-5P DNA barcode data","year":2021,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Environmental DNA in Biodiversity Studies","field":"Environmental Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Guelph","funders":"Ontario Ministry of Economic Development, Job Creation and Trade; Ministero dello Sviluppo Economico; Ontario Genomics; Compute Canada; Genome Canada","keywords":"Barcode; Indel; DNA barcoding; Biology; DNA sequencing; Computational biology; Hidden Markov model; Sequence (biology); In silico; Computer science; Data mining; Artificial intelligence; Pattern recognition (psychology); Genetics; DNA; Evolutionary biology; Gene","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":[0.007400754,0.002058425,0.001581769,0.002677454,0.001365819,0.002642354,0.002520284,0.001652329,0.0094308],"category_scores_gemma":[0.0178402,0.001614087,0.001936373,0.001613993,0.001176807,0.001689892,0.00266029,0.004961727,0.01223292],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008620318,"about_ca_system_score_gemma":0.001503592,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002085896,"about_ca_topic_score_gemma":0.004632006,"domain_scores_codex":[0.9963294,0.0007986213,0.0003433556,0.001096681,0.001222621,0.0002093633],"domain_scores_gemma":[0.9961995,0.001618777,0.0005346648,0.0007111926,0.0008130281,0.0001228003],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002082216,0.0004199447,0.009935953,0.002896659,0.001296537,0.000952032,0.001992532,0.0238806,0.1859176,0.01984885,0.1366265,0.6141505],"study_design_scores_gemma":[0.0003368118,0.0004035748,0.01070082,0.0003959811,0.0002928214,0.0009363675,0.0003658589,0.3934443,0.3153129,0.02114609,0.2561016,0.0005630429],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0100164,0.0005412472,0.8734459,0.0002225683,0.0003313736,0.0002872102,0.006018508,0.1076971,0.001439677],"genre_scores_gemma":[0.01886804,0.0001995443,0.9538116,0.0002936534,0.00004878668,0.0005330471,0.008319502,0.01535911,0.002566736],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0094308,"threshold_uncertainty_score":0.03913939,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0456696989857359,"score_gpt":0.2425646635699268,"score_spread":0.1968949645841909,"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."}}