Structure-based drug design by denoising voxel grids

TL;DR

VoxBind introduces a novel 3D voxel-denoising generative model for structure-based drug design, enabling faster, more diverse molecule generation with improved binding affinity predictions.

Contribution

It extends the neural empirical Bayes formalism to conditional 3D molecule generation, offering a simpler, faster, and more effective approach for structure-based drug design.

Findings

More diverse generated molecules

Fewer steric clashes in molecules

Higher binding affinity to protein pockets

Abstract

We present VoxBind, a new score-based generative model for 3D molecules conditioned on protein structures. Our approach represents molecules as 3D atomic density grids and leverages a 3D voxel-denoising network for learning and generation. We extend the neural empirical Bayes formalism (Saremi & Hyvarinen, 2019) to the conditional setting and generate structure-conditioned molecules with a two-step procedure: (i) sample noisy molecules from the Gaussian-smoothed conditional distribution with underdamped Langevin MCMC using the learned score function and (ii) estimate clean molecules from the noisy samples with single-step denoising. Compared to the current state of the art, our model is simpler to train, significantly faster to sample from, and achieves better results on extensive in silico benchmarks -- the generated molecules are more diverse, exhibit fewer steric clashes, and bind with higher affinity to protein pockets. The code is available at https://github.com/genentech/voxbind/.