{"id":"W2156995746","doi":"10.1186/1471-2105-10-316","title":"Unsupervised statistical clustering of environmental shotgun sequences","year":2009,"lang":"en","type":"article","venue":"BMC Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":111,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University","funders":"Advanced Research Projects Agency; Defense Advanced Research Projects Agency; Burroughs Wellcome Fund","keywords":"Cluster analysis; Computer science; Shotgun sequencing; Dimensionality reduction; Unsupervised learning; Pairwise comparison; Statistical model; Pattern recognition (psychology); Artificial intelligence; Computational biology; Data mining; Genome; Biology; Genetics","routes":{"ca_aff":true,"ca_fund":false,"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":[0.001256708,0.0005162797,0.0005749322,0.00207442,0.0004602638,0.000672028,0.001258392,0.0005854186,0.0009261479],"category_scores_gemma":[0.003898073,0.0003145663,0.0007141989,0.001596724,0.0009442538,0.0007996244,0.0007438794,0.0007430072,0.0005251589],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008282164,"about_ca_system_score_gemma":0.001106651,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001358505,"about_ca_topic_score_gemma":0.002290365,"domain_scores_codex":[0.9989109,0.0002521116,0.00007741003,0.0003371698,0.0003636156,0.0000587357],"domain_scores_gemma":[0.9973933,0.00112963,0.0004658596,0.0003664626,0.0005601454,0.00008468334],"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":[0.000403783,0.0002658664,0.02593221,0.001083326,0.0004148026,0.0003879198,0.0005636118,0.4506139,0.2120364,0.03779804,0.003380099,0.26712],"study_design_scores_gemma":[0.00001679159,0.00005950874,0.009911973,0.00004844338,0.00002499838,0.0002505345,0.00007915124,0.8990543,0.04859447,0.03841215,0.003491867,0.00005583226],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06474987,0.0001886801,0.9324574,0.00006956203,0.00001318021,0.00007247926,0.0005239465,0.001314906,0.0006099481],"genre_scores_gemma":[0.2906423,0.000251723,0.7047099,0.0000726244,0.00002790829,0.0002000964,0.002907079,0.0003870209,0.0008012772],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00207442,"threshold_uncertainty_score":0.006646156,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01362830568214565,"score_gpt":0.229514811453265,"score_spread":0.2158865057711194,"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."}}