Abstract Glioblastoma (GBM) is an aggressive brain tumour with limited responsiveness to current immunotherapeutic approaches, partly due to its low mutational burden and intra-tumour heterogeneity. A systematic understanding of the tumour antigen landscape is therefore essential for advancing tumour immunology and supporting rational development of immunotherapeutic strategies. In this study, we performed whole-transcriptome sequencing of RNA extracted from 79 formalin-fixed paraffin-embedded (FFPE) IDH-wildtype GBM samples to systematically identify and prioritise candidate tumour antigens derived from three sources: single-nucleotide variants (SNVs), overexpressed tumour-associated antigens (TAAs), and gene fusion events. Candidate peptides were evaluated using integrated computational criteria, including transcript expression, predicted antigen processing features, peptide–HLA binding affinity and stability. Across the cohort, mutation-derived tumor-specific antigens (TSAs) were largely private to individual samples, whereas TAAs constituted a larger and more recurrent candidate pool. Despite comparable predicted binding characteristics across antigen classes, recurrence patterns differed substantially, reflecting their distinct biological origins. Fusion-derived candidates were rare and sample-specific. Predicted peptide presentation was disproportionately associated with a limited subset of HLA class I alleles. Collectively, this study provides a systematically prioritized catalogue of transcriptionally expressed GBM antigen candidates and offers a comparative evaluation of mutation-, expression-, and fusion-derived antigen sources within a unified transcriptome-based framework.
Kert et al. (Tue,) studied this question.