Application of Machine Learning in mRNA Therapeutics Development

Abstract Messenger RNA (mRNA) therapeutics enable rapid and programmable protein expression in vaccines and disease treatment, and have emerged as a revolutionary breakthrough in modern medicine. However, due to the complex interplay among the composition of mRNA sequence, structural stability, translation efficiency, and intracellular delivery, the design of mRNA molecules remains challenging. Recent advances in machine learning (ML), particularly deep learning, have greatly accelerated mRNA research by enabling data-driven modeling. In this review, we summarized the application of ML in mRNA therapeutics, including prediction of RNA secondary and tertiary structure, sequence optimization, translation efficiency modeling, splicing prediction, epitranscriptomic modification prediction, and delivery system design. The key ML methodologies underlying these advances were also introduced. Finally, we discuss existing challenges, including data scarcity, poor model interpretability, and insufficient experimental validation, and outline future directions for building fully integrated artificial intelligence–driven mRNA drug development pipelines.