R-CCS Cafe 第229回 第1部

Although molecular generation models based on deep learning have attracted significant attention in drug discovery, most existing models have a limitation in the context of drug design wherein they do not sufficiently consider the effect of the three-dimensional (3D) structure of the target protein in the generation process. In this talk, we introduce a novel de novo molecular design method, SBMolGen, which uses the docking score as a reward function to generate high-affinity score compounds for a specified target. More specifically, this deep learning-based molecular generator integrates a recurrent neural network, Monte Carlo tree search, and docking simulations. SBMolGen not only generates novel binding active molecules but also presents 3D docking poses with target proteins, which will be useful in subsequent drug design. The code is available at https://github.com/clinfo/SBMolGen.