{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T11:10:32Z","timestamp":1788952232132,"version":"build-2803163510"},"reference-count":45,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T00:00:00Z","timestamp":1788912000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Immunol."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>\n                      Peptide therapeutics are increasingly explored to treat challenging diseases, but immunogenicity risks limit their clinical success.\n                      <jats:italic>In silico<\/jats:italic>\n                      tools enable immunogenicity screening through prediction of peptide-MHCII binding, yet current methods do not account for chemical properties of non-natural amino acids routinely incorporated to improve peptide properties.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>Here, we present a machine learning approach combining chemical fingerprints with sequence information to predict MHC class II binding for peptides including both natural (NAA) and non-natural (NNAA) amino acids. We evaluate peptide-level fingerprints against residue-level fingerprints (direct-encoding and similarity-based fingerprints) that preserve positional information while encoding chemical diversity, for a total of 31 different representations.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Peptide-level fingerprints showed poor performance due to loss of positional information, while residue-level fingerprints matched the performance of sequence-based encodings (BLOSUM62 and one-hot) for NAA-based peptides while accurately identifying binding cores and motifs. Considering citrullinated peptides as a case study, we observed similar linear correlation performance across encoding strategies, with residue-level fingerprints showing marginal improvements in quantitative prediction accuracy. Citrulline was rarely found at canonical anchor positions, suggesting that the advantage of explicit chemical encoding may be more pronounced for modifications occurring at positions critical for binding.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>The proposed framework allows for inclusion of diverse modifications, including synthetic amino acids, and is compatible with existing pan-allele architectures. While the full advantage of explicit chemical encoding remains to be demonstrated, this framework is designed to capture these effects as experimental data becomes available, supporting immunogenicity risk assessment for emerging peptide therapeutics.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fimmu.2026.1869623","type":"journal-article","created":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T10:16:51Z","timestamp":1788949011000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Beyond natural amino acids: extending immunogenicity risk assessment to modified peptides through chemical feature encoding. A case study on citrullination"],"prefix":"10.3389","volume":"17","author":[{"given":"Maria","family":"Cairoli","sequence":"first","affiliation":[{"name":"Discovery Computation and AI, Hit Discovery, Discovery Sciences, Biopharmaceuticals R&D, AstraZeneca","place":["Gothenburg, Sweden"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Morten","family":"Nielsen","sequence":"additional","affiliation":[{"name":"Department of Health Technology, Technical University of Denmark","place":["Lyngby, Denmark"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Catherine","family":"Betts","sequence":"additional","affiliation":[{"name":"Immune Safety, Clinical Pharmacology and Safety Sciences, Biopharmaceuticals R&D, AstraZeneca","place":["Cambridge, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Olga","family":"Obrezanova","sequence":"additional","affiliation":[{"name":"Biologics Engineering, Oncology R&D, AstraZeneca","place":["Cambridge, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Leonardo","family":"De Maria","sequence":"additional","affiliation":[{"name":"Discovery Computation and AI, Hit Discovery, Discovery Sciences, Biopharmaceuticals R&D, AstraZeneca","place":["Gothenburg, Sweden"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1038\/s41573-020-00135-8","article-title":"Trends in peptide drug discovery","volume":"20","author":"Muttenthaler","year":"2021","journal-title":"Nat Rev Drug Discov"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.2174\/138161210793292555","article-title":"Chemical modifications designed to improve peptide stability: Incorporation of non-natural amino acids, pseudo-peptide bonds, and cyclization","volume":"16","author":"Gentilucci","year":"2010","journal-title":"Curr Pharm Des"},{"key":"B3","doi-asserted-by":"publisher","first-page":"389","DOI":"10.1038\/s41573-020-0062-z","article-title":"Advances in therapeutic peptides targeting G protein-coupled receptors","volume":"19","author":"Davenport","year":"2020","journal-title":"Nat Rev Drug Discov"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1021\/acsmedchemlett.3c00037","article-title":"Beyond 20 in the 21st century: Prospects and challenges of non-canonical amino acids in peptide drug discovery","volume":"14","author":"Hickey","year":"2023","journal-title":"ACS Med Chem Lett"},{"key":"B5","doi-asserted-by":"publisher","first-page":"898","DOI":"10.1038\/s41573-024-01051-x","article-title":"Immunogenicity risk assessment and mitigation for engineered antibody and protein therapeutics","volume":"23","author":"Carter","year":"2024","journal-title":"Nat Rev Drug Discov"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.1016\/j.bmcl.2014.10.016","article-title":"Amino acidic scaffolds bearing unnatural side chains: An old idea generates new and versatile tools for the life sciences","volume":"24","author":"Stevenazzi","year":"2014","journal-title":"Bioorg Med Chem Lett"},{"key":"B7","doi-asserted-by":"publisher","first-page":"2999","DOI":"10.1021\/acs.jcim.2c00193","article-title":"Virtual screening expands the non-natural amino acid palette for peptide optimization","volume":"62","author":"Amarasinghe","year":"2022","journal-title":"J Chem Inf Model"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.3389\/fimmu.2021.637963","article-title":"Introduction of non-natural amino acids into T-cell epitopes to mitigate peptide-specific T-cell responses","volume":"12","author":"Azam","year":"2021","journal-title":"Front Immunol"},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.3389\/fddsv.2022.952326","article-title":"In silico immunogenicity assessment for sequences containing unnatural amino acids: A method using existing in silico algorithm infrastructure and a vision for future enhancements","volume":"2","author":"Mattei","year":"2022","journal-title":"Front Drug Discov"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1016\/j.drudis.2023.103714","article-title":"Immunogenicity risk assessment of synthetic peptide drugs and their impurities","volume":"28","author":"De Groot","year":"2023","journal-title":"Drug Discov Today"},{"key":"B11","doi-asserted-by":"publisher","first-page":"191","DOI":"10.1146\/annurev-biodatasci-021920-100259","article-title":"Immunoinformatics: Predicting peptide-MHC binding","volume":"3","author":"Nielsen","year":"2020","journal-title":"Annu Rev BioMed Data Sci"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.1186\/1471-2105-8-238","article-title":"Prediction of MHC class II binding affinity using SMM-align, a novel stabilization matrix alignment method","volume":"8","author":"Nielsen","year":"2007","journal-title":"BMC Bioinf"},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.1186\/1471-2105-10-296","article-title":"NN-align. An artificial neural network-based alignment algorithm for MHC class II peptide binding prediction","volume":"10","author":"Nielsen","year":"2009","journal-title":"BMC Bioinf"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkaa379","article-title":"NetMHCpan-4.1 and NetMHCIIpan-4.0: Improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data","volume":"48","author":"Reynisson","year":"2020","journal-title":"Nucleic Acids Res"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1126\/scitranslmed.aag1286","article-title":"Post hoc assessment of the immunogenicity of bioengineered factor VIIa demonstrates the use of preclinical tools","volume":"9","author":"Lamberth","year":"2017","journal-title":"Sci Transl Med"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1021\/acs.jproteome.5c01284","article-title":"NetMHCIIphosPan: A machine learning tool for predicting HLA class II antigen presentation of phosphorylated peptides","volume":"25","author":"Alvarez","year":"2026","journal-title":"J Proteome Res"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1021\/ci100050t","article-title":"Extended-connectivity fingerprints","volume":"50","author":"Rogers","year":"2010","journal-title":"J Chem Inf Model"},{"key":"B18","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1186\/s13321-024-00849-6","article-title":"One chiral fingerprint to find them all","volume":"16","author":"Orsi","year":"2024","journal-title":"J Cheminf"},{"key":"B19","doi-asserted-by":"publisher","first-page":"185","DOI":"10.1186\/s13321-025-01115-z","article-title":"How to build machine learning models able to extrapolate from standard to modified peptides","volume":"17","author":"Fernandez-Diaz","year":"2025","journal-title":"J Cheminform"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.1039\/d4dd00056k","article-title":"A methodology to correctly assess the applicability domain of cell membrane permeability predictors for cyclic peptides","volume":"3","author":"Geylan","year":"2024","journal-title":"Digital Discov"},{"key":"B21","doi-asserted-by":"publisher","first-page":"992","DOI":"10.1038\/s41557-021-00766-3","article-title":"Deep learning to design nuclear-targeting abiotic miniproteins","volume":"13","author":"Schissel","year":"2021","journal-title":"Nat Chem"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkae1092","article-title":"The immune epitope database (IEDB): 2024 update","volume":"53","author":"Vita","year":"2025","journal-title":"Nucleic Acids Res"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1002\/pro.5560010313","article-title":"Selection of representative protein data sets","volume":"1","author":"Hobohm","year":"1992","journal-title":"Protein Sci"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkae1010","article-title":"UniProt: The universal protein knowledgebase in 2025","volume":"53","author":"The UniProt Consortium","year":"2025","journal-title":"Nucleic Acids Res"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.1038\/s41594-024-01214-9","article-title":"A quantitative and site-specific atlas of the citrullinome reveals widespread existence of citrullination and insights into PADI4 substrates","volume":"31","author":"Rebak","year":"2024","journal-title":"Nat Struct Mol Biol"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.89.22.10915","article-title":"Amino acid substitution matrices from protein blocks","volume":"89","author":"Henikoff","year":"1992","journal-title":"Proc Natl Acad Sci"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/24.1.197","article-title":"The Blocks Database-a system for protein classification","volume":"24","author":"Pietrokovski","year":"1996","journal-title":"Nucleic Acids Res"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0067844","article-title":"An index for characterization of natural and non-natural amino acids for peptidomimetics","volume":"8","author":"Liang","year":"2013","journal-title":"PloS One"},{"key":"B29","doi-asserted-by":"publisher","first-page":"35","DOI":"10.1186\/s13321-024-00830-3","article-title":"Effectiveness of molecular fingerprints for exploring the chemical space of natural products","volume":"16","author":"Boldini","year":"2024","journal-title":"J Cheminf"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1002\/psc.3666","article-title":"PepFunn: Novo Nordisk open-source toolkit to enable peptide in silico analysis","volume":"31","author":"Ochoa","year":"2025","journal-title":"J Pept Sci"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1021\/acs.biochem.3c00253","article-title":"AmorProt: Amino acid molecular fingerprints repurposing-based protein fingerprint","volume":"62","author":"Lee","year":"2023","journal-title":"Biochemistry"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.1021\/ci010132r","article-title":"Reoptimization of MDL keys for use in drug discovery","volume":"42","author":"Durant","year":"2002","journal-title":"J Chem Inf Comput Sci"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1711.05101","article-title":"Decoupled weight decay regularization","author":"Loshchilov","year":"2017","journal-title":"arXiv"},{"key":"B34","doi-asserted-by":"publisher","first-page":"84","DOI":"10.1186\/s13073-018-0594-6","article-title":"Footprints of antigen processing boost MHC class II natural ligand predictions","volume":"10","author":"Barra","year":"2018","journal-title":"Genome Med"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/bth100","article-title":"Improved prediction of MHC class I and class II epitopes using a novel Gibbs sampling approach","volume":"20","author":"Nielsen","year":"2004","journal-title":"Bioinformatics"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1145\/3292500.3330701","article-title":"Optuna: A next-generation hyperparameter optimization framework","author":"Akiba","year":"2019","journal-title":"KDD '19: Proc 25th ACM SIGKDD Int Conf Knowledge Discov Data Minin"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gks469","article-title":"Seq2Logo: A method for construction and visualization of amino acid binding motifs and sequence profiles including sequence weighting, pseudo counts and two-sided representation of amino acid enrichment and depletion","volume":"40","author":"Thomsen","year":"2012","journal-title":"Nucleic Acids Res"},{"key":"B38","doi-asserted-by":"publisher","DOI":"10.1038\/9858","article-title":"Generation of tissue-specific and promiscuous HLA ligand databases using DNA microarrays and virtual HLA class II matrices","volume":"17","author":"Sturniolo","year":"1999","journal-title":"Nat Biotechnol"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.3389\/fimmu.2017.00292","article-title":"Major histocompatibility complex (MHC) class I and MHC class II proteins: Conformational plasticity in antigen presentation","volume":"8","author":"Wieczorek","year":"2017","journal-title":"Front Immunol"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkac965","article-title":"The MHC motif atlas: A database of MHC binding specificities and ligands","volume":"51","author":"Tadros","year":"2023","journal-title":"Nucleic Acids Res"},{"key":"B41","doi-asserted-by":"publisher","DOI":"10.4049\/jimmunol.160.7.3363","article-title":"Several common HLA-DR types share largely overlapping peptide binding repertoires","volume":"160","author":"Southwood","year":"1998","journal-title":"J Immunol"},{"key":"B42","doi-asserted-by":"publisher","DOI":"10.4049\/jimmunol.151.11.5930","article-title":"HLA-A1 and HLA-A3 T cell epitopes derived from influenza virus proteins predicted from peptide binding motifs","volume":"151","author":"DiBrino","year":"1993","journal-title":"J Immunol"},{"key":"B43","doi-asserted-by":"publisher","DOI":"10.4049\/jimmunol.157.8.3480","article-title":"Specificity and degeneracy in peptide binding to HLA-B7-like class I molecules","volume":"157","author":"Sidney","year":"1996","journal-title":"J Immunol"},{"key":"B44","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0177140","article-title":"Citrullination only infrequently impacts peptide binding to HLA class II MHC","volume":"12","author":"Sidney","year":"2017","journal-title":"PloS One"},{"key":"B45","doi-asserted-by":"publisher","DOI":"10.64898\/2026.03.12.711352","article-title":"Chemically informed representations of amino acids enable learning beyond the canonical protein alphabet","author":"Christiansen","year":"2026","journal-title":"BiorXiv"}],"container-title":["Frontiers in Immunology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fimmu.2026.1869623\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T10:16:52Z","timestamp":1788949012000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fimmu.2026.1869623\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,9]]},"references-count":45,"alternative-id":["10.3389\/fimmu.2026.1869623"],"URL":"https:\/\/doi.org\/10.3389\/fimmu.2026.1869623","relation":{},"ISSN":["1664-3224"],"issn-type":[{"value":"1664-3224","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,9]]},"article-number":"1869623"}}