Unified all-atom molecule generation with neural fields
This work addresses the problem of limited modality in generative models for structure-based drug design, offering a unified approach that could benefit drug discovery, though it appears incremental as it adapts existing computer vision methods.
The authors tackled the challenge of generating target-conditioned, all-atom molecules across diverse atomic systems by introducing FuncBind, a framework using neural fields and score-based generative models, which achieved competitive in silico performance for small molecules, macrocyclic peptides, and antibody loops, and generated novel antibody binders in vitro.
Generative models for structure-based drug design are often limited to a specific modality, restricting their broader applicability. To address this challenge, we introduce FuncBind, a framework based on computer vision to generate target-conditioned, all-atom molecules across atomic systems. FuncBind uses neural fields to represent molecules as continuous atomic densities and employs score-based generative models with modern architectures adapted from the computer vision literature. This modality-agnostic representation allows a single unified model to be trained on diverse atomic systems, from small to large molecules, and handle variable atom/residue counts, including non-canonical amino acids. FuncBind achieves competitive in silico performance in generating small molecules, macrocyclic peptides, and antibody complementarity-determining region loops, conditioned on target structures. FuncBind also generated in vitro novel antibody binders via de novo redesign of the complementarity-determining region H3 loop of two chosen co-crystal structures. As a final contribution, we introduce a new dataset and benchmark for structure-conditioned macrocyclic peptide generation. The code is available at https://github.com/prescient-design/funcbind.