{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001776652,0.0001312338,0.0002549676,0.00004410528,0.0002466503,0.00000783909,0.0005375188,0.00005591253,0.00006510728],"category_scores_gemma":[0.000008803201,0.0001064306,0.00005616832,0.0001185661,0.0004275393,0.00008789066,0.0007732002,0.0001252262,9.720729e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002770541,"about_ca_system_score_gemma":0.00002668511,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001502757,"about_ca_topic_score_gemma":0.01449989,"domain_scores_codex":[0.9991424,0.0002019661,0.0002629596,0.0001214542,0.0001210151,0.0001501712],"domain_scores_gemma":[0.998996,0.0001977637,0.0001335268,0.0006314295,0.00001255947,0.0000287701],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003339611,0.007378193,0.597058,0.0002363899,0.002911047,0.00002674247,0.06875319,0.001861605,0.2928811,0.003611586,0.005085286,0.01685723],"study_design_scores_gemma":[0.02750077,0.0106144,0.118855,0.0003473672,0.001413754,0.0007499614,0.1777671,0.003046249,0.3184946,0.007129655,0.3312117,0.002869472],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9928191,0.004649481,8.386079e-7,0.0003715011,0.00006869261,0.0005474425,0.001410988,0.000008579682,0.0001233258],"genre_scores_gemma":[0.9853001,0.0008319192,0.01256185,0.00001608442,0.000007847219,0.0003317145,0.0009185576,0.000005960223,0.00002600767],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.478203,"threshold_uncertainty_score":0.8091277,"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."}}