(334a) AI-Aided Design of Biomaterials for mRNA Delivery

The deployment of mRNA vaccines, especially for combating SARS-CoV-2, has brought to the forefront the indispensable role of ionizable lipid nanoparticles (LNPs) in the realm of mRNA therapeutics. Despite their proven efficacy, the broader application of mRNA therapies faces significant challenges due to the need for LNPs that are specifically tailored for various cellular targets. The conventional pathway to developing these LNPs is marked by intensive labor, high costs, and a predominantly trial-and-error methodology, which collectively stymie the advancement of mRNA-based treatments. In response to these challenges, we propose an innovative solution that leverages artificial intelligence (AI) to redefine the process of designing ionizable lipids, which are key to the successful delivery of mRNA via LNPs. Our approach integrates AI with principles of combinatorial chemistry to enhance the development process, characterized by the rapid generation of diverse lipid libraries, the use of deep learning models for the predictive screening of lipids, and the adaptability to a range of cell types. Employing this AI-driven strategy, we have efficiently screened over 10,000 ionizable lipid candidates, pinpointing those with optimal characteristics for targeted delivery to muscle and immune cells. This process not only revealed a clear preference for certain lipid attributes depending on the cell type but also emphasized the necessity for variability in lipid structures, such as tail lengths and head groups, to achieve maximal delivery efficacy. The outcomes of this research pave the way for a new era in mRNA therapy development, offering a scalable and precise method for creating customized LNPs. By addressing the intricate requirements of mRNA delivery, our AI-aided approach promises to significantly broaden the therapeutic potential of mRNA technologies.