{"id":"W4408586797","doi":"10.1093/protein/gzaf003","title":"Tuning ProteinMPNN to reduce protein visibility via MHC Class I through direct preference optimization","year":2025,"lang":"en","type":"article","venue":"Protein Engineering Design and Selection","topic":"vaccines and immunoinformatics approaches","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Asian Studies Association; University of Victoria","funders":"Infrastruktura PL-Grid; Fundacja na rzecz Nauki Polskiej; European Commission; Royal Academy of Engineering; European Regional Development Fund; Horizon 2020 Framework Programme; UK Research and Innovation; Intelligence Community Postdoctoral Research Fellowship Program; Israel Cancer Research Fund","keywords":"MHC class I; Computer science; Visibility; Major histocompatibility complex; Epitope; Workflow; Computational biology; Immune system; Antigen; Biology; Genetics; Database","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003389795,0.0001961557,0.0001554323,0.00008018011,0.0001271023,0.00007741079,0.0001064866,0.0001456503,0.000003611695],"category_scores_gemma":[0.0001903875,0.0001940895,0.00003472334,0.0003195555,0.00001145514,0.00002733829,0.00006977234,0.00012776,0.000001408051],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004326785,"about_ca_system_score_gemma":0.00006178761,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002160906,"about_ca_topic_score_gemma":0.00000149955,"domain_scores_codex":[0.9990206,0.00005378249,0.0002509463,0.0003316044,0.00008811231,0.0002550036],"domain_scores_gemma":[0.999606,0.000006979664,0.00005649996,0.000177838,0.0001017647,0.00005091969],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001217745,0.00003057207,0.00001221774,0.0001385322,0.00003411013,8.478163e-8,0.00006300006,0.190411,0.8061826,0.0001161607,0.00004614571,0.002843832],"study_design_scores_gemma":[0.0002760592,0.0002975808,0.00009486818,0.0001083701,0.00001059741,0.000003152029,0.000009876354,0.3390725,0.658361,0.00006356604,0.001505241,0.0001971662],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1034167,0.0002372093,0.8938792,0.0001038146,0.00003823728,0.001985607,9.675399e-7,0.00006783141,0.0002704462],"genre_scores_gemma":[0.8460412,0.00002002623,0.1525477,0.00003252539,0.00005001636,0.000704524,0.00001694695,0.00002012967,0.0005669147],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7426246,"threshold_uncertainty_score":0.791474,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01543434391653663,"score_gpt":0.2233008051834782,"score_spread":0.2078664612669415,"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."}}