Researchers at MIT developed COMET, a transformer-based machine learning model, to optimize lipid nanoparticle (LNP) formulations for RNA vaccine and therapy delivery. By analyzing thousands of particles, COMET predicts novel LNPs with enhanced cell-targeting and efficacy. Experimental validation confirmed improved protein expression both in vitro and in vivo. This AI-accelerated design approach promises to streamline development of RNA medicines for diseases like obesity and diabetes, marking a significant advance in nucleic acid therapeutic platforms.