Learning a Continuous Representation of 3D Molecular Structures with Deep Generative Models
Machine learning methods in drug discovery have primarily focused on virtual
screening of molecular libraries using discriminative models. Generative models
are an entirely different approach to drug discovery that learn to represent
and optimize molecules in a continuous latent space. These methods have already
been applied with increasing success to the generation of two dimensional
molecules as SMILES strings and molecular graphs. In this work, we describe
deep generative models for three dimensional molecular structures using atomic
density grids and a novel fitting algorithm that converts continuous grids to
discrete molecular structures. Our models jointly represent drug-like molecules
and their conformations in a latent space that can be explored through
interpolation. We are able to sample diverse sets of molecules based on a given
input compound and increase the probability of creating a valid, drug-like
molecule.