From prediction to precision: how immunopeptidomics advances neoantigen discovery

Abstract T cell recognition of peptides presented by class I and II human leukocyte antigen (HLA) molecules is fundamental to cancer immunity and personalized immunotherapy. Neoantigens, peptides containing somatic mutations, are attractive therapeutic targets due to their tumor specificity and immunogenicity. Current neoantigen discovery pipelines rely on sequencing and computational predictions but often struggle to identify peptides that are both presented and immunogenic. Immunopeptidomics, which uses mass spectrometry to identify naturally presented HLA-bound peptides, enables detection of neoantigens displayed on tumor cells. This review explores how immunopeptidomics complements existing tools to refine neoantigen identification, improves machine learning prediction algorithms through immunopeptidomics-derived datasets, and distinguishes between mutant and wild-type immunopeptides. We also highlight emerging developments that will further integrate immunopeptidomics into personalized immunotherapy.