
AI system RhoAgent designs functional RNA with 6x efficiency gains
Researchers have introduced RhoAgent, an autonomous AI system capable of designing functional RNA molecules by iteratively refining candidates using experimental feedback. The system leverages a diffusion-based generative tool called RDiff to program three-dimensional structures, overcoming limitations in current computational methods. Validated through multiple benchmarks, RhoAgent generated active ribozymes with 40–70% sequence similarity to wild-type counterparts and improved the cleavage efficiency of the hovlinc ribozyme by up to sixfold. Additionally, it created a fluorogenic aptamer scaffold that increased fluorescence by 230% and designed 3D scaffolds for Förster resonance energy transfer systems, preserving the binding activities of distinct functional cores. This advancement represents a significant step in synthetic biology and therapeutic development, enabling the precise engineering of RNA for complex biological functions.