{"id":"W2078685372","doi":"10.1142/9789812702456_0035","title":"DISCOVERING SEQUENCE-STRUCTURE MOTIFS FROM PROTEIN SEGMENTS AND TWO APPLICATIONS","year":2004,"lang":"en","type":"article","venue":"","topic":"Machine Learning in Bioinformatics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Cluster analysis; Computer science; Sequence (biology); Protein secondary structure; Cluster (spacecraft); Protein structure prediction; Structural alignment; Local structure; Protein tertiary structure; Support vector machine; Protein structure; Artificial intelligence; Dynamic programming; Data structure; Data mining; Pattern recognition (psychology); Sequence alignment; Algorithm; Peptide sequence; Biology; Physics; 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":[0.0003495252,0.0005664326,0.0004235164,0.0007474479,0.0004120613,0.0004404763,0.0005806393,0.0009700985,0.001149975],"category_scores_gemma":[0.001802101,0.0002292257,0.0003067431,0.001248803,0.0005006686,0.0005569232,0.0005259239,0.0006165369,0.0005736658],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001978142,"about_ca_system_score_gemma":0.0002858617,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001191315,"about_ca_topic_score_gemma":0.001404321,"domain_scores_codex":[0.9997696,0.00005797047,0.00001474868,0.0000750258,0.00005908442,0.00002350066],"domain_scores_gemma":[0.9993157,0.0002950943,0.00006634257,0.00008944165,0.0001566019,0.00007678549],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001143974,0.0007234626,0.01308625,0.0003070277,0.00008387196,0.001133343,0.0004183792,0.1097669,0.3051657,0.01414164,0.002122947,0.5519065],"study_design_scores_gemma":[0.0001035163,0.0004522954,0.006536447,0.00001827868,0.00002803502,0.000904078,0.0001465714,0.8518316,0.1182453,0.01820054,0.00349181,0.00004150814],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.4501184,0.0009179771,0.545278,0.0004829718,0.00002985996,0.00009565672,0.0002400598,0.000909881,0.001927301],"genre_scores_gemma":[0.6273306,0.0003201905,0.3680524,0.0000701531,0.0000445703,0.0000945149,0.0007454186,0.00007117865,0.003271008],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.001191315,"threshold_uncertainty_score":0.003847063,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.005945630023237146,"score_gpt":0.2548494911316115,"score_spread":0.2489038611083744,"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."}}