Manipulating 3D Molecules in a Fixed-Dimensional E(3)-Equivariant Latent Space

TL;DR

This paper introduces MolFLAE, a novel E(3)-equivariant VAE for 3D molecules that enables zero-shot manipulation and optimization of molecular structures in a fixed-dimensional latent space.

Contribution

The work presents MolFLAE, a fixed-dimensional, E(3)-equivariant VAE for 3D molecules that supports zero-shot editing and optimization, independent of atom counts.

Findings

Achieves competitive results on 3D molecule generation benchmarks.

Enables zero-shot molecule manipulation such as atom editing and structure interpolation.

Demonstrates improved drug properties while maintaining key interactions.

Abstract

Medicinal chemists often optimize drugs considering their 3D structures and designing structurally distinct molecules that retain key features, such as shapes, pharmacophores, or chemical properties. Previous deep learning approaches address this through supervised tasks like molecule inpainting or property-guided optimization. In this work, we propose a flexible zero-shot molecule manipulation method by navigating in a shared latent space of 3D molecules. We introduce a Variational AutoEncoder (VAE) for 3D molecules, named MolFLAE, which learns a fixed-dimensional, E(3)-equivariant latent space independent of atom counts. MolFLAE encodes 3D molecules using an E(3)-equivariant neural network into fixed number of latent nodes, distinguished by learned embeddings. The latent space is regularized, and molecular structures are reconstructed via a Bayesian Flow Network (BFN) conditioned on the encoder's latent output. MolFLAE achieves competitive performance on standard unconditional 3D molecule generation benchmarks. Moreover, the latent space of MolFLAE enables zero-shot molecule manipulation, including atom number editing, structure reconstruction, and coordinated latent interpolation for both structure and properties. We further demonstrate our approach on a drug optimization task for the human glucocorticoid receptor, generating molecules with improved hydrophilicity while preserving key interactions, under computational evaluations. These results highlight the flexibility, robustness, and real-world utility of our method, opening new avenues for molecule editing and optimization.