{"id":"W2079257324","doi":"10.1109/cibcb.2010.5510344","title":"Issues with the PipeAlign phylogenomics toolkit in identifying protein subfamilies","year":2010,"lang":"en","type":"article","venue":"","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University","funders":"","keywords":"Phylogenomics; Phylogenetic tree; Biology; Consistency (knowledge bases); Computational biology; Phylogenetics; Annotation; Evolutionary biology; Protein sequencing; Similarity (geometry); Bioinformatics; Gene; Computer science; Genetics; Artificial intelligence; Clade; Peptide sequence","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":[0.07297235,0.002622245,0.002235797,0.00353444,0.003466264,0.007250813,0.009580592,0.002737303,0.004342893],"category_scores_gemma":[0.1509209,0.003724452,0.002543995,0.006264806,0.003736728,0.01007717,0.007763562,0.008605173,0.009344795],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001302327,"about_ca_system_score_gemma":0.00339384,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00483106,"about_ca_topic_score_gemma":0.007044234,"domain_scores_codex":[0.9523623,0.02289186,0.004451614,0.004394833,0.01477796,0.001121409],"domain_scores_gemma":[0.8457214,0.09978446,0.006745681,0.0269486,0.0183054,0.00249447],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00649573,0.001229022,0.06211866,0.006742703,0.001487848,0.002485211,0.009731286,0.04495504,0.08767892,0.04966155,0.1350692,0.5923448],"study_design_scores_gemma":[0.0008640718,0.001809079,0.03686004,0.002335653,0.0006383214,0.008099484,0.004176916,0.3671572,0.1235758,0.1369736,0.3159633,0.001546504],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07560194,0.003920068,0.8336862,0.00848639,0.001088692,0.0006992141,0.002939308,0.06649458,0.007083588],"genre_scores_gemma":[0.09458078,0.001494213,0.8777481,0.002533926,0.0002851793,0.0006678306,0.004780726,0.0150271,0.002882077],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.07297235,"threshold_uncertainty_score":0.3859193,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009313124105133217,"score_gpt":0.234044119408395,"score_spread":0.2247309953032618,"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."}}