Background and aims: There is a growing interest in the application of messenger ribonucleicacid (mRNA) vaccines as an alternative to conventional vaccine approaches due to their low manufacturingcosts, rapid development, safe administration, and high potency. However, unstableand inefficient in vivo delivery of mRNA has hindered their application. Moreover, the first stepin developing a vaccine is to identify possible antigens. To discover and optimize new effectivevaccine candidates, AI-based models have shown promising results.In this review, we provide a systematic review of the development of mRNA vaccines throughthe application of AI and how AI is particularly valuable in the research and production of mRNAvaccines.Methods: We searched PubMed, Scopus, or Web of Science for data till March ۲۰۲۳ for publishedstudies of AI applications on the development of mRNA vaccines. The Systematic Reviewschecklist was applied. Keywords: [(Artificial intelligence OR
machine learning OR deeplearning OR
neural network OR
random forest OR support vector machine) and (mRNA vaccineor messenger ribonucleic acid vaccine) and (medicine or drug)]. We used PROBAST (predictionmodel risk of bias assessment tool) to assess the quality of literature related to the Safety and Immunogenicityof mRNA Vaccines. The inclusion criteria for paper selection were: ۱) Paper mustbe peer-reviewed. ۲) Journals on which papers are published must be either PubMed, Scopus, orWeb of Science indexed. ۳) The paper should use only AI techniques. Exclusion criteria for paperselection were: ۱) Duplicate studies in different databases. ۲) Study which is less cited by otherpeer-reviewed papers. ۳) MSc and Ph.D. papers.Results: From the ۱۰۷ identified records [PubMed® (n = ۴۰); Scopus (n = ۳۳) and or Web ofScience (n = ۳۴)], ۲۸ studies were included. Among others, the selected studies on new mRNAvaccines were classified, as follows: studies with in-vivo and; studies with in-vivo and/or clinicaldata; and other studies related to mRNA vaccines. This review provided sufficient evidence to delineatethe potential of AI in analyzing
mRNA vaccine features for mRNA modeling on multipleaspects like evaluating the solubility and other physicochemical parameters, Immuno-informaticsanalyses, and molecular docking analysis. Different methods were identified, mainly from thearea of machine learning. The most used techniques were support vector machine, random forest,and artificial
neural network models.Conclusion: Diverse potential mRNA vaccines were identified. AI was a suitable tool to quicklyanalyze large amounts of data or to develop mRNA vaccines.