{"id":"W2920549681","doi":"10.1093/bioinformatics/btz151","title":"VHost-Classifier: virus-host classification using natural language processing","year":2019,"lang":"en","type":"article","venue":"Bioinformatics","topic":"Bacteriophages and microbial interactions","field":"Environmental Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Classifier (UML); Computer science; Artificial intelligence; Python (programming language); Virus classification; Identifier; Host (biology); Biology; Programming language; Genetics; Genome","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":[0.002538251,0.002244256,0.001135064,0.003280595,0.001187828,0.002639773,0.002497444,0.001758711,0.01295904],"category_scores_gemma":[0.007226679,0.0005891717,0.001772029,0.002022605,0.0006010755,0.002397231,0.001811588,0.001756251,0.01628123],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001368644,"about_ca_system_score_gemma":0.002611902,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005110305,"about_ca_topic_score_gemma":0.007053004,"domain_scores_codex":[0.9976084,0.0003594168,0.0002557182,0.0008145251,0.0006744341,0.0002875314],"domain_scores_gemma":[0.9975611,0.001177748,0.0002005789,0.0002476854,0.0006667168,0.000146202],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001313397,0.0007750164,0.02318085,0.002998029,0.0004462547,0.0005546776,0.0004528076,0.01717964,0.02362519,0.004400092,0.4967664,0.4283076],"study_design_scores_gemma":[0.0004101398,0.0006699974,0.01540926,0.0005447983,0.0002156035,0.001549831,0.0006709702,0.6899943,0.06643678,0.02599017,0.1978129,0.0002953314],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"software","genre_gemma":"empirical","genre_scores_codex":[0.0855398,0.003157537,0.3398494,0.002395084,0.001647119,0.002450345,0.1669846,0.3763849,0.02159117],"genre_scores_gemma":[0.1811083,0.000825196,0.5324616,0.001582154,0.0004795843,0.001875819,0.2605105,0.007755985,0.01340092],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01295904,"threshold_uncertainty_score":0.04335231,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01561674675428714,"score_gpt":0.2596760937011227,"score_spread":0.2440593469468356,"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."}}