{"id":"W4406905918","doi":"10.1128/mbio.02650-24","title":"Machine learning reveals the dynamic importance of accessory sequences for <i>Salmonella</i> outbreak clustering","year":2025,"lang":"en","type":"article","venue":"mBio","topic":"Salmonella and Campylobacter epidemiology","field":"Agricultural and Biological Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"Mitacs; Genome British Columbia; Genome Canada","keywords":"Genome; Computational biology; Bacterial genome size; Cluster analysis; Outbreak; Biology; Genetics; Genotyping; Data mining; Computer science; Artificial intelligence; Gene; Virology; Genotype","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006161672,0.0001223708,0.0002583564,0.00001010585,0.0002038944,0.00002276316,0.0003986857,0.00009152495,0.00005038676],"category_scores_gemma":[0.0001854902,0.00004097853,0.0001244841,0.0002231558,0.0001008578,0.00004929997,0.0001127779,0.0001248198,0.000003340156],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001870227,"about_ca_system_score_gemma":0.000007537607,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002334416,"about_ca_topic_score_gemma":0.001022107,"domain_scores_codex":[0.9989522,0.0001272233,0.0003373968,0.0002581269,0.00007090075,0.0002541316],"domain_scores_gemma":[0.9988821,0.0007706713,0.0001891761,0.00007900303,0.00005035569,0.0000286709],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00008850677,0.0000359141,0.6931306,0.00005716215,0.00004779255,0.000001624475,0.0000857515,0.0001218999,0.2711014,0.0006538222,0.0006456529,0.03402988],"study_design_scores_gemma":[0.000456997,0.0004208036,0.9128611,0.0001544229,0.00008851173,0.00001665931,0.0009029862,0.007053915,0.006874107,0.01939636,0.05131368,0.0004604347],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.991859,0.001944717,0.0001139067,0.003537091,0.0001989872,0.0002617476,0.00003610743,0.00004777471,0.00200069],"genre_scores_gemma":[0.9968689,0.0002125922,0.00009186634,0.0009169046,0.0000549777,0.00003473032,0.00004497319,9.520166e-7,0.001774066],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2642273,"threshold_uncertainty_score":0.1671056,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02523374509289805,"score_gpt":0.2757108712652237,"score_spread":0.2504771261723256,"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."}}