Abstract Background: Despite significant progress in therapeutic outcomes, melanoma remains frequently resistant to immunotherapy. Efforts to improve response rates include the development of vaccines targeting tumor neoantigens; however, the optimal tumor-associated antigens (TAAs) to target remain uncertain. Recent evidence suggests that the most relevant TAAs may originate from non-mutated protein sequences, including those derived from the so-called dark proteome (1). In this study, we conducted an immunopeptidomic analysis of tumor samples obtained from patients with metastatic melanoma before initiation of systemic therapy Methods: A total of 35 tumor biopsy samples from melanoma patients were subjected to native tissue lysis, followed by immunoaffinity purification of HLA class I complexes using a pan HLA class I antibody (W6/32) and subsequent elution of the bound immunopeptides. The isolated immunopeptides were analyzed using data-independent acquisition (DIA) mass spectrometry on an Exploris 480 mass spectrometer (Thermo Scientific) equipped with FAIMS Pro. Data analysis was performed using Spectronaut 20 (Biognosys) using directDIA. For peptide identification, a customized human proteome database was constructed to enable the detection of tumor-associated antigens (TAs). This database included the canonical and isoforms human proteome, a previously reported TA reference dataset (1), and an in-house generated “dark genome” database derived from tumor RNA-sequencing data. Results: Analysis of these 35 melanoma tissue samples yielded in total over 120,000 unique immunopeptides with an average of ∼15,000 per sample, with notable inter-sample variability (ranging from ∼6,000 to 28,000 identifications). All samples displayed the expected HLA class I length distribution, with a predominant population of 9-mer peptides. Across the cohort, approx. 70 previously reported tumor-associated antigens (TAs) were identified, spanning multiple categories including tumor-associated antigens (TAA), lineage-specific antigens (LSA), and aberrantly expressed tumor-specific antigens (aeTSA). These TAs originated from both annotated open reading frames (ORFs) and noncanonical translation events such as frameshifts, noncoding RNA (ncRNA), and 5′UTR-derived peptides. From the RNA-seq-derived database, approximately 22 immunopeptides were mapped to sequences originating from both annotated genes (18 peptides) and previously uncharacterized genomic regions (4 peptides; dark genome). Together, these results demonstrate that the applied proteogenomic immunopeptidomics workflow enables the detection of tumor-specific and neoantigenic peptides directly from clinical melanoma biopsy samples, underscoring its potential utility in translational immuno-oncology research. Reference: (1) Apavaloaei et al. Nature Cancer, 2025 Citation Format: Arthur Viodé, Daniel Gautheret, Anamarija Pfeiffer, Hughes Hermann, Sharane Muralli, Séverine Roy, Antoine Meant, Amaury Lachaud, Yuehan Feng, Caroline Robert. Discovery of tumor-specific antigens in melanoma tissue biopsies via integrated proteogenomic analysis abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 5729.
Viodé et al. (Fri,) studied this question.