{"id":"W4379012186","doi":"10.26434/chemrxiv-2023-pqjrf","title":"Accurately predicting protein pKa values using non-equilibrium alchemy","year":2023,"lang":"en","type":"preprint","venue":"ChemRxiv","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Government of Ontario","keywords":"Protonation; Chemistry; Lysine; Thermodynamics; Physics; Amino acid; Biochemistry","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001476755,0.0005543861,0.0005899652,0.0003171826,0.0001644808,0.0007879141,0.002673998,0.0003694317,0.0000115461],"category_scores_gemma":[0.0006011466,0.000602439,0.0003349272,0.0007746505,0.0001062731,0.0006990329,0.006785669,0.0009558228,0.0001119868],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002742209,"about_ca_system_score_gemma":0.0009597415,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008969373,"about_ca_topic_score_gemma":0.000001171686,"domain_scores_codex":[0.9960066,0.0002343769,0.0007138241,0.001557068,0.0008273439,0.0006607978],"domain_scores_gemma":[0.997022,0.0004723733,0.0005398884,0.001487441,0.0002798614,0.0001984321],"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":[0.00003675889,0.000287145,0.001824282,0.002214128,0.0005484864,0.0002825689,0.004131783,0.8248667,0.151584,0.003457964,0.002245081,0.008521123],"study_design_scores_gemma":[0.000192285,0.00001669272,0.001531575,0.0005944411,0.00002494729,0.00001206319,0.00002829949,0.8561732,0.09704652,0.04372978,0.00007596047,0.0005741999],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5296945,0.00009444993,0.4662082,0.0004413,0.001579973,0.0006759153,0.000005605824,0.0006212729,0.0006787999],"genre_scores_gemma":[0.6430289,0.000006589119,0.3546246,0.0001122426,0.001119987,0.0001934276,0.00005238839,0.000125798,0.0007360748],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1133344,"threshold_uncertainty_score":0.9996427,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1423244323327653,"score_gpt":0.3751456542771446,"score_spread":0.2328212219443793,"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."}}