Peptide-binding motifs define the functional architecture of human leukocyte antigens

Background Peptide presentation by polymorphic human leukocyte antigens (HLA) is central to adaptive and innate immunity. HLA polymorphism gives rise to allele-specific peptide-binding motifs (PBM) that shape the immunopeptidomes presented across individuals. However, the global functional architecture imposed by this diversity remains poorly understood. Methods Here, we leveraged large-scale immunopeptidomics datasets to perform a comprehensive analysis of PBM across HLA class I (HLA-A, -B, -C) and class II (HLA-DR, -DQ, -DP) molecules. PBM were derived from public resources and de novo motif deconvolution of 334,089 class II peptides, followed by hierarchical clustering to define functional groups. We systematically evaluated the association of these groups with immunopeptidome overlap and their relation to established models of HLA functional diversity. Results We identified 40 discrete PBM-defined functional groups across 155 class I and 93 class II allotypes. Molecules within the same PBM group showed significantly higher immunopeptidome similarity compared to those across groups. In class II, PBM grouping was further shaped by HLA-DRB3/4/5 paralogues and by α-chain polymorphism in HLA-DQ and -DP, which modulated group assignment of specific β-chains. PBM classification showed only partial concordance with serological antigen groups and evolutionary divergence metrics, but captured differences in peptide repertoires with higher resolution than existing sequence- or epitope-based models. Overall, PBM-based grouping provided a more direct representation of immunopeptidome diversity. Conclusions This pan-HLA PBM framework comprehensively maps the functional architecture of antigen presentation and provides a unified platform to functionally classify HLA diversity. By capturing immunopeptidome divergence, PBM groups offer a foundation for improving the interpretation of HLA variation in immunogenetics, transplantation, and immune-mediated diseases.