From Target to Drug: Generative Modeling for the Multimodal Structure-Based Ligand Design

Mol Pharm. 2019 Oct 7;16(10):4282-4291. doi: 10.1021/acs.molpharmaceut.9b00634. Epub 2019 Sep 10.

Abstract

Chemical space is impractically large, and conventional structure-based virtual screening techniques cannot be used to simply search through the entire space to discover effective bioactive molecules. To address this shortcoming, we propose a generative adversarial network to generate, rather than search, diverse three-dimensional ligand shapes complementary to the pocket. Furthermore, we show that the generated molecule shapes can be decoded using a shape-captioning network into a sequence of SMILES enabling directly the structure-based de novo drug design. We evaluate the quality of the method by both structure- (docking) and ligand-based [quantitative structure-activity relationship (QSAR)] virtual screening methods. For both evaluation approaches, we observed enrichment compared to random sampling from initial chemical space of ZINC drug-like compounds.

Keywords: deep learning; generative modeling; structure based drug design.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Drug Design*
  • Drug Discovery*
  • Humans
  • Ligands
  • Models, Chemical*
  • Molecular Conformation
  • Neural Networks, Computer*
  • Proteins / chemistry*
  • Proteins / metabolism
  • Quantitative Structure-Activity Relationship
  • Small Molecule Libraries / chemistry*
  • Small Molecule Libraries / metabolism

Substances

  • Ligands
  • Proteins
  • Small Molecule Libraries