Peptide vaccines have emerged as a versatile platform complementing traditional vaccines by offering high safety, precise epitope targeting, and ease of manufacturing; however, they suffer from intrinsically weak immunogenicity, human leukocyte antigen (HLA) restriction, and poor in vivo stability. Recent progress in immunoinformatics and artificial intelligence (AI) has transformed epitope discovery from empirical trial-and-error screening toward rational, systems-level design. Machine learning models trained on binding assays, liquid chromatography–tandem mass spectrometry–derived ligandomes, and T-cell response data now enable increasingly accurate prediction of peptide–major histocompatibility complex binding, T-cell receptor recognition, and even the integrated antigen-presentation cascade, spanning sequence-based, structure-informed, and pan-allelic frameworks. In parallel, medicinal chemistry strategies─including cyclization, lipidation, glycosylation, stapling, incorporation of noncanonical amino acids, and construction of multicomponent or tolerogenic conjugates─address key biopharmaceutical bottlenecks by enhancing immunogenicity, extending half-life, and enabling organ- and cell-type–specific delivery. Together, these computational and chemical advances are beginning to bridge the gap between in silico performance and clinical efficacy, supporting the development of peptide vaccines that can be tailored to diverse HLA backgrounds, disease settings, and therapeutic goals. This review summarizes current progress at the interface of AI-driven epitope prediction and chemical modification with a focus on how their integration can yield next-generation peptide vaccines with improved potency, durability, and safety.
Wang et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: