{"id":"W4409333289","doi":"10.1128/mbio.00699-25","title":"Eukfinder: a pipeline to retrieve microbial eukaryote genome sequences from metagenomic data","year":2025,"lang":"en","type":"article","venue":"mBio","topic":"Microbial Community Ecology and Physiology","field":"Environmental Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"Canadian Institutes of Health Research","keywords":"Metagenomics; Genome; Biology; Computational biology; Microbiome; Human Microbiome Project; Bacterial genome size; Sequence assembly; Shotgun sequencing; Genomics; Genetics; Gene; Transcriptome","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":[0.003113676,0.003141033,0.001551684,0.005943528,0.001387106,0.002306403,0.003545643,0.001637721,0.007621928],"category_scores_gemma":[0.004599916,0.002066378,0.002224636,0.002798519,0.0006532011,0.002954232,0.00350863,0.001909133,0.005608778],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008088748,"about_ca_system_score_gemma":0.002137965,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001994542,"about_ca_topic_score_gemma":0.002356178,"domain_scores_codex":[0.9985589,0.0002202223,0.0001531661,0.0005763872,0.0003547654,0.0001365053],"domain_scores_gemma":[0.998811,0.0004045072,0.0002433005,0.0001996514,0.0001956892,0.0001457887],"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":[0.004086186,0.0006996567,0.01493238,0.007061641,0.001509852,0.00227444,0.001636586,0.02423776,0.3962238,0.01001606,0.15291,0.3844115],"study_design_scores_gemma":[0.001026593,0.0006095096,0.01718447,0.0007622843,0.0005074884,0.001719225,0.0005921293,0.4334924,0.3048359,0.02731127,0.2111391,0.0008196781],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.0266423,0.00170207,0.6485841,0.0004522397,0.0001917947,0.0007060796,0.05762998,0.2616127,0.002478765],"genre_scores_gemma":[0.07124159,0.0008087161,0.8147448,0.0004670471,0.00007111027,0.001457586,0.09724993,0.011491,0.002468274],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.007621928,"threshold_uncertainty_score":0.02549785,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0252722375234503,"score_gpt":0.2667538463827485,"score_spread":0.2414816088592981,"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."}}