Advances and Challenges in De Novo Drug Design Using Three-Dimensional Deep Generative Models

A persistent goal forde novodrug design is togenerate novel chemical compounds with desirable properties in alabor-, time-, and cost-efficient manner. Deep generative modelsprovide alternative routes to this goal. Numerous modelarchitectures and optimization strategies have been explored inrecent years, most of which have been developed to generate two-dimensional molecular structures. Some generative models aimingat three-dimensional (3D) molecule generation have also beenproposed, gaining attention for their unique advantages andpotential to directly design drug-like molecules in a target-conditioning manner. This review highlights current developmentsin 3D molecular generative models combined with deep learningand discusses future directions forde novodrug design.