{"id":"W2059926522","doi":"10.1016/j.bbrc.2007.02.040","title":"Prediction of protein crystallization using collocation of amino acid pairs","year":2007,"lang":"en","type":"article","venue":"Biochemical and Biophysical Research Communications","topic":"Machine Learning in Bioinformatics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":118,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"In silico; Protein crystallization; Naive Bayes classifier; Computer science; Crystallization; Classifier (UML); Sequence (biology); Protein Data Bank (RCSB PDB); Artificial intelligence; Threading (protein sequence); Protein sequencing; Peptide sequence; Computational biology; Pattern recognition (psychology); Protein structure; Algorithm; Chemistry; Biology; Biochemistry; Support vector machine","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009089924,0.0009109952,0.001261353,0.002748121,0.000760539,0.0007250766,0.0006401672,0.0009992189,0.001080809],"category_scores_gemma":[0.003553323,0.0005412329,0.001088908,0.00204612,0.0004381561,0.001152967,0.0007669064,0.0009558895,0.0005584687],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005662485,"about_ca_system_score_gemma":0.001068942,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003449685,"about_ca_topic_score_gemma":0.005891742,"domain_scores_codex":[0.9994909,0.0001340015,0.00005567535,0.0001434528,0.0001311874,0.00004484319],"domain_scores_gemma":[0.9975647,0.001061204,0.0003824033,0.0003134478,0.0004997288,0.0001785236],"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":[0.006448364,0.00082707,0.0993347,0.0009030786,0.000951724,0.00247569,0.0003238615,0.4541868,0.1705949,0.007313057,0.0126178,0.244023],"study_design_scores_gemma":[0.00006202198,0.00008379513,0.003603573,0.00001138381,0.00006025628,0.0001675589,0.00002672217,0.9805619,0.01233852,0.002502577,0.0005667973,0.00001479786],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.767209,0.001076594,0.2250864,0.0002301878,0.0001225592,0.000157879,0.001817985,0.002764473,0.001534953],"genre_scores_gemma":[0.9299107,0.000231969,0.06724901,0.00003806836,0.00002396636,0.0000503117,0.002076172,0.0001283865,0.0002913859],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003449685,"threshold_uncertainty_score":0.006859183,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07206198932686944,"score_gpt":0.3613267705409309,"score_spread":0.2892647812140614,"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."}}