AI and the World's Hard Problems · Medicine & drug discovery · arXiv · 2026-10-19
Generative molecule evolution using 3D pharmacophore for efficient Structure-Based Drug Design
Structure-based drug design faces scarce protein-ligand complex data and the question of whether generated molecules can be synthesized. The authors propose MEVO, which evolves molecules within a synthesizable chemical space using a reaction-aware generative model trained on reaction routes built from Enamine building blocks, guided by a hierarchical score that includes ABFEP calculations. They report that MEVO designed potent inhibitors for the challenging target KRAS G12D, with affinities comparable to known active inhibitors.