{"id":"W4378783088","doi":"10.1101/gr.277395.122","title":"Genealogical inference and more flexible sequence clustering using iterative-PopPUNK","year":2023,"lang":"en","type":"article","venue":"Genome Research","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"European and Developing Countries Clinical Trials Partnership; Youth Innovation Promotion Association; National Key Research and Development Program of China; Shanghai Rising-Star Program; Youth Innovation Promotion Association of the Chinese Academy of Sciences; Department for International Development; Medical Research Council; Chinese Academy of Sciences; Medical Research Council Canada; National Natural Science Foundation of China; European Commission","keywords":"Biology; Inference; Cluster analysis; Genome; Annotation; Population; Iterative method; Computational biology; Bacterial genome size; Data mining; Computer science; Genetics; Artificial intelligence; Algorithm; Gene","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007538089,0.001358312,0.001596898,0.002660587,0.002266249,0.002185829,0.00354314,0.001675751,0.003640749],"category_scores_gemma":[0.02692856,0.00118742,0.002467741,0.002219949,0.001301763,0.002537578,0.003191658,0.002513916,0.001765175],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001192041,"about_ca_system_score_gemma":0.00215493,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004475904,"about_ca_topic_score_gemma":0.008963491,"domain_scores_codex":[0.9970644,0.001472242,0.0001738481,0.0006800463,0.0004510981,0.0001583844],"domain_scores_gemma":[0.9896477,0.00624933,0.0005595688,0.002329861,0.0009719996,0.0002416192],"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":[0.001035681,0.0004511308,0.02847016,0.0008041175,0.001147482,0.0007792492,0.002365223,0.622761,0.02701893,0.04824415,0.008722601,0.2582003],"study_design_scores_gemma":[0.00005252628,0.00003969121,0.001182926,0.00003369356,0.00004133847,0.0001514783,0.0001096824,0.9645606,0.006425254,0.02436294,0.002983823,0.00005599008],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04401997,0.00009871362,0.9484383,0.0001302114,0.00003749548,0.00008560844,0.0005071702,0.0056757,0.001006745],"genre_scores_gemma":[0.1374575,0.00005223789,0.8576158,0.0001186386,0.00001792069,0.0001767758,0.001880962,0.002054256,0.0006259496],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007538089,"threshold_uncertainty_score":0.03986573,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2548564996911687,"score_gpt":0.4422091171988944,"score_spread":0.1873526175077257,"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."}}