Background: Epitope-based cancer vaccines require rigorous safety filtering and structural validation to overcome translational limitations in heterogeneous malignancies such as gastric cancer.
To design and computationally validate a safety-oriented multi-epitope vaccine targeting gastric cancer–associated membrane proteins.
Methods: Extracellular domains of nine gastric cancer–associated membrane proteins were screened to identify CTL (9-mer) and HTL (15-mer) epitopes. Peptides with exact sequence matches to the human reference proteome were excluded as a preliminary self-similarity filtering step. Epitope prioritization was performed using NetMHCpan and NetMHCIIpan based on binding affinity and percentile rank. Selected candidates were evaluated for antigenicity, allergenicity, toxicity, proteasomal cleavage, and TAP transport efficiency. High-ranking epitopes were assembled using optimized linkers with an N-terminal β-defensin-3 adjuvant. Physicochemical properties, intrinsic disorder, and secondary structure were analyzed. The tertiary structure was predicted, refined, and validated using Ramachandran plot and TM-score metrics. Vaccine–HLA interactions were assessed by molecular docking and molecular dynamics simulations. Immune responses were simulated in silico, and codon optimization with in silico cloning was performed to evaluate expression feasibility.
Results: Screening of extracellular domains from nine gastric cancer–associated membrane proteins generated 161–900 unique CTL and 109–894 HTL non-self peptides per antigen. Six epitopes were prioritized based on low percentile ranks (CTL ≤2; HTL ≤3.1) and favorable antigenicity (VaxiJen 0.51–0.99). The final 143–amino-acid construct showed a molecular weight of 15.54 kDa, instability index of 27.02, and GRAVY score of −0.435, indicating predicted stability and hydrophilicity. Structural refinement yielded 97.2% residues in favored/allowed Ramachandran regions (90.4% core) with RMSD 0.48 Å and TM-score 0.9886, confirming preservation of global topology. ClusPro docking identified the top-ranked clusters for both complexes, with weighted scores of −881.7 (HLA-A*02:01) and −878.3 (HLA-DR1). Molecular dynamics simulations showed rapid equilibration (~10 ns), stable RMSD, consistent radius of gyration, gradual SASA reduction, and persistent hydrogen bonds. Codon optimization achieved a CAI of 1.0 and GC content of 50.82%, supporting expression feasibility.
Conclusion: This integrative strategy provides a computationally prioritized and structurally validated multi-epitope vaccine candidate. Nevertheless, the applied filtering strategy represents only a preliminary risk-reduction approach, and all findings require experimental validation.