Score-based 3D molecule generation with neural fields

TL;DR

This paper presents FuncMol, a novel neural field-based model for unconditional 3D molecule generation that encodes molecular structures as continuous atomic density fields, enabling scalable, fast, and structure-agnostic molecule synthesis.

Contribution

The paper introduces a new neural field representation for 3D molecules and a walk-jump sampling method, enabling efficient, unconditional generation of diverse molecules without structural assumptions.

Findings

Achieves competitive results on drug-like molecules

Scales well with molecular size, including macro-cyclic peptides

Samples molecules at least ten times faster than previous methods

Abstract

We introduce a new representation for 3D molecules based on their continuous atomic density fields. Using this representation, we propose a new model based on walk-jump sampling for unconditional 3D molecule generation in the continuous space using neural fields. Our model, FuncMol, encodes molecular fields into latent codes using a conditional neural field, samples noisy codes from a Gaussian-smoothed distribution with Langevin MCMC (walk), denoises these samples in a single step (jump), and finally decodes them into molecular fields. FuncMol performs all-atom generation of 3D molecules without assumptions on the molecular structure and scales well with the size of molecules, unlike most approaches. Our method achieves competitive results on drug-like molecules and easily scales to macro-cyclic peptides, with at least one order of magnitude faster sampling. The code is available at https://github.com/prescient-design/funcmol.