{"id":"W6898783637","doi":"10.57745/irrmxh","title":"Taxonomic or trait assignation of environmental sequences in regions with a poor taxonomic knowledge: case of river diatom metabarcoding with a new version of the annotated reference library Diat.barcode - supplementary data","year":2025,"lang":"en","type":"dataset","venue":"Recherche Data Gouv France","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Continental (Canada)","funders":"","keywords":"Trait; DNA barcoding; Mainland; Diatom; Taxonomic rank; Biodiversity","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":[0.00353289,0.00119641,0.00119175,0.004352876,0.0009629444,0.001704705,0.001412933,0.001213494,0.009277449],"category_scores_gemma":[0.008998482,0.0003564542,0.000910635,0.004909031,0.0003962732,0.0006502573,0.001410658,0.0009954975,0.007751395],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001147003,"about_ca_system_score_gemma":0.00152244,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02685131,"about_ca_topic_score_gemma":0.05323401,"domain_scores_codex":[0.9966625,0.0007007702,0.0003998815,0.001483295,0.0004424198,0.0003110457],"domain_scores_gemma":[0.9959205,0.001618085,0.0005009643,0.001089774,0.0006946988,0.0001759505],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.002577511,0.0002448883,0.1576849,0.01275935,0.001930826,0.001572418,0.002076027,0.00644727,0.021729,0.004393468,0.6787763,0.1098081],"study_design_scores_gemma":[0.0002795179,0.0001202427,0.2142032,0.00119747,0.0006385027,0.0009197442,0.0006933439,0.004117515,0.005303272,0.002115042,0.7702765,0.0001356832],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.0355643,0.001337985,0.002968989,0.0002230013,0.00008049688,0.00005389415,0.9559717,0.001318918,0.002480649],"genre_scores_gemma":[0.01683339,0.0001332368,0.004751068,0.00008095408,0.00001142372,0.0001008888,0.9772259,0.0001903298,0.0006728263],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.02685131,"threshold_uncertainty_score":0.05339003,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1898644262286296,"score_gpt":0.343977544889338,"score_spread":0.1541131186607084,"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."}}