{"id":"W3138735990","doi":"10.1111/1755-0998.13384","title":"Debar: A sequence‐by‐sequence denoiser for COI‐5P DNA barcode data","year":2021,"lang":"en","type":"article","venue":"Molecular Ecology Resources","topic":"Environmental DNA in Biodiversity Studies","field":"Environmental Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Guelph","funders":"Ontario Ministry of Economic Development, Job Creation and Trade","keywords":"Barcode; Indel; Biology; DNA barcoding; In silico; Computational biology; Hidden Markov model; Sequence (biology); DNA sequencing; Word error rate; Genetics; Evolutionary biology; Artificial intelligence; Computer science; DNA; Gene; Single-nucleotide polymorphism","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":[0.007132222,0.002376781,0.001764738,0.003123409,0.001519415,0.002758261,0.002738838,0.001663463,0.01006104],"category_scores_gemma":[0.01827771,0.001865928,0.002442291,0.001867091,0.001179931,0.001890917,0.003228736,0.005694599,0.01492627],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009177693,"about_ca_system_score_gemma":0.001643403,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002204228,"about_ca_topic_score_gemma":0.005609693,"domain_scores_codex":[0.9963211,0.0008355975,0.0003405117,0.001210831,0.001061751,0.0002301304],"domain_scores_gemma":[0.9964052,0.0015516,0.0005777476,0.0006219667,0.0007198418,0.0001235999],"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":[0.00210438,0.0003789726,0.01209508,0.003382465,0.001487277,0.0009811332,0.002850455,0.01752835,0.1604536,0.01595367,0.1266524,0.6561322],"study_design_scores_gemma":[0.000393001,0.0004970714,0.01505542,0.0005850353,0.0004436481,0.001286256,0.000654171,0.3129349,0.2626574,0.02395323,0.380765,0.0007748592],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01223553,0.0008410954,0.8739365,0.0002687301,0.0004099901,0.0003754144,0.008863838,0.1011519,0.001916933],"genre_scores_gemma":[0.019375,0.0003154832,0.9501351,0.0003847162,0.00006165687,0.0006709,0.01064109,0.01576103,0.002654976],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01006104,"threshold_uncertainty_score":0.03771925,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03547885858086927,"score_gpt":0.2606308441238946,"score_spread":0.2251519855430254,"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."}}