Authors: Nomathamsanqa Tholo1, Gavin Markey1, Ruairidh Harrigan1, Preeti Pandey2, Bodhayan Prasad3, Ram Shankar Barai4, David Samuel Gibson1, Priyank Shukla1

Affiliations: 1. Personalised Medicine Centre, School of Medicine, Ulster University, C-TRIC Building, Altnagelvin Area Hospital, Glenshane Road, Londonderry, BT47 6SB, UK. 2. Department of Genetics & Biochemistry, Clemson University, 190 Collings St., Clemson, SC 29634, USA. 3. Wolfson Wohl Cancer Research Centre, School of Cancer Sciences, University of Glasgow (Garscube Campus), Glasgow G61 1QH, UK. 4. Biological Sciences Division, ICMR – National Institute of Occupational Health, Meghani Nagar, Ahmedabad 380016, Gujarat, India.

Background/ Introduction: Peptide-based vaccines have emerged as a promising strategy for rapid, safe, and scalable responses to emerging infectious diseases. Advances in artificial intelligence (AI) and bioinformatics have transformed vaccine design from empirical approaches to computationally driven pipelines. However, variability in prediction accuracy, reproducibility, and global applicability remains a challenge. This literature review aimed to critically evaluate how AI and bioinformatics are transforming peptide vaccine discovery and to identify the key scientific and translational gaps that must be addressed.

Material & Methods: A comprehensive literature review was conducted, synthesising evidence from over a decade of published immunoinformatic pipelines (2015-2025), epitope prediction tool evaluations, and clinical case studies in both infectious diseases and oncology. Studies were analysed comparatively based on pipeline architecture, prediction tools, validation strategies, and translational outcomes. Emphasis was placed on consensus epitope prediction, HLA population coverage, structural validation, and emerging AI-driven approaches.

Results: The review helped in constructing a comprehensive computational pipeline architecture for designing peptide-based vaccine candidates. Steps of this pipeline involve antigen selection, multi-tool epitope prediction, consensus scoring, safety filtering, population coverage analysis, structural modelling, docking, and immune simulation. AI-enabled tools, including pan-allelic MHC predictors and deep learning models have significantly improved prediction performance and scalability. However, key challenges persist, including limited performance in conformational B-cell epitope prediction, dataset bias, lack of standardised reproducibility frameworks, and underrepresentation of diverse HLA populations. Evidence from infectious disease and oncology studies demonstrated that peptide vaccines are increasingly becoming clinically feasible when combined with advanced formulation and delivery systems.

Conclusion: AI and bioinformatics are reshaping and strengthening pandemic preparedness capabilities by accelerating peptide vaccine discovery, enabling rapid response to global health threats. Future progress will depend on improving reproducibility, explainability, equitable data representation, and integration with experimental validation. These advances will be critical for translating peptide-based vaccines into scalable and globally deployable precision medicine platforms.