{"id":"W4225700116","doi":"10.1038/s41467-022-29438-7","title":"MiDAS 4: A global catalogue of full-length 16S rRNA gene sequences and taxonomy for studies of bacterial communities in wastewater treatment plants","year":2022,"lang":"en","type":"article","venue":"Nature Communications","topic":"Wastewater Treatment and Nitrogen Removal","field":"Environmental Science","cited_by":386,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Villum Fonden","keywords":"16S ribosomal RNA; Taxonomy (biology); Gene; Computational biology; Ribosomal RNA; Biology; Genetics; Ecology","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":[0.002767741,0.003164025,0.001960165,0.009556343,0.001272153,0.001857668,0.001668072,0.001446759,0.003181901],"category_scores_gemma":[0.005054092,0.0009282004,0.001850036,0.008156245,0.0003966799,0.001852872,0.002423953,0.001435139,0.005617077],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007185243,"about_ca_system_score_gemma":0.002765466,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002030812,"about_ca_topic_score_gemma":0.004574904,"domain_scores_codex":[0.9974083,0.0003853363,0.000417421,0.0007627695,0.0007186436,0.0003075317],"domain_scores_gemma":[0.9968734,0.000445824,0.001146093,0.0004708786,0.0007184856,0.0003452902],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.004181374,0.0005246676,0.07269695,0.01257925,0.001168264,0.0008668903,0.001528978,0.005966471,0.5288853,0.002759533,0.08517136,0.283671],"study_design_scores_gemma":[0.0003353391,0.001166882,0.1941686,0.001726987,0.0009612424,0.001952571,0.0008093462,0.01039561,0.1368004,0.004475136,0.6468379,0.0003699468],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.1446337,0.01033988,0.07261232,0.0002196255,0.0003156458,0.0007264108,0.7487906,0.01427786,0.008083845],"genre_scores_gemma":[0.04590896,0.001899593,0.1093285,0.0001383624,0.00006014119,0.0009205091,0.8387609,0.001317733,0.001665261],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.009556343,"threshold_uncertainty_score":0.01463741,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06041746253887677,"score_gpt":0.2927521953921114,"score_spread":0.2323347328532346,"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."}}