mRNAutilus: Multi-Objective-Guided Discrete Generation of mRNA with Optimized Therapeutic Properties
Researchers have developed mRNAutilus, a novel framework for designing therapeutic mRNA sequences. This system uses a masked discrete diffusion model combined with Monte Carlo Tree Guidance to optimize multiple functional objectives simultaneously, including codon usage and untranslated regions. Unlike previous methods that design these components separately, mRNAutilus generates complete transcripts in a single, optimized process. The framework has demonstrated significant improvements in protein expression for various applications, outperforming existing commercial and machine learning-designed constructs. AI
IMPACT Enables more efficient and effective design of therapeutic mRNA, potentially accelerating drug development.