{"id":"W4206020930","doi":"10.1101/2022.01.14.476062","title":"Optimized phylogenetic clustering of HIV-1 sequence data for public health applications","year":2022,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"HIV Research and Treatment","field":"Immunology and Microbiology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Western University","funders":"Center for AIDS Research, University of Washington","keywords":"Statistics; Covariate; Poisson distribution; Phylogenetic tree; Cluster analysis; Null model; Poisson regression; Range (aeronautics); Biology; Population; Sampling (signal processing); Mathematics; Econometrics; Demography; Computer science; Ecology; Genetics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.0106743,0.001016397,0.001186915,0.003077792,0.001208103,0.001345562,0.001446844,0.00115776,0.002345959],"category_scores_gemma":[0.03445847,0.0007786619,0.001285626,0.002136205,0.0006426387,0.001403806,0.001614448,0.001695584,0.0008306638],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001448842,"about_ca_system_score_gemma":0.001868564,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005421042,"about_ca_topic_score_gemma":0.007609894,"domain_scores_codex":[0.9948649,0.003572366,0.000239943,0.0007508777,0.0003972086,0.0001748099],"domain_scores_gemma":[0.9844024,0.009755812,0.001345298,0.001955512,0.002220396,0.0003206016],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001368868,0.0006457411,0.06009709,0.0006321514,0.000872814,0.0002672267,0.001671601,0.655706,0.01819221,0.01032993,0.0102689,0.2399475],"study_design_scores_gemma":[0.00003127682,0.0000734256,0.004124905,0.00002514424,0.00002768659,0.0000452299,0.0001336675,0.9865639,0.002748127,0.005267766,0.0009387871,0.00001997639],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2790853,0.0003325655,0.7096786,0.0005201047,0.00006808659,0.0004958554,0.001721129,0.006870645,0.00122774],"genre_scores_gemma":[0.3637324,0.00007076542,0.6325406,0.00007608503,0.00001978247,0.0002613504,0.002238764,0.0006422501,0.000417938],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0106743,"threshold_uncertainty_score":0.05645174,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07801090410801755,"score_gpt":0.3057499958502585,"score_spread":0.227739091742241,"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."}}