MuCO: Generative Peptide Cyclization Empowered by Multi-stage Conformation Optimization
Researchers have developed MuCO, a novel generative method for peptide cyclization that improves the modeling of diverse conformations. The method employs a three-stage process: topology-aware backbone design, generative side-chain packing, and physics-aware all-atom optimization. This approach allows for efficient exploration of low-energy conformations and has demonstrated superior performance on the CPSea dataset compared to existing methods in terms of stability, diversity, and efficiency. AI
IMPACT This method could accelerate the discovery and design of new cyclic peptides for pharmaceutical applications.