SiDGen: Structure-informed Diffusion for Generative modeling of Ligands for Proteins

TL;DR

SiDGen introduces a structure-informed diffusion model for ligand generation that balances interaction detail and computational efficiency, enabling scalable and accurate structure-based drug design.

Contribution

It proposes a novel Topological Information Bottleneck to compress protein representations, reducing computational costs while maintaining accuracy.

Findings

Achieves state-of-the-art results on key benchmarks.

Reduces memory and computational costs significantly.

Bridges the gap between sequence efficiency and structural detail.

Abstract

Structure-based drug design (SBDD) faces a fundamental scaling fidelity dilemma: rich pocket-aware conditioning captures interaction geometry but can be costly, often scales quadratically ($O(L^2)$) or worse with protein length ($L$), while efficient sequence-only conditioning can miss key interaction structure. We propose SiDGen, a structure-informed discrete diffusion framework that resolves this trade-off through a Topological Information Bottleneck (TIB). SiDGen leverages a learned, soft assignment mechanism to compress residue-level protein representations into a compact bottleneck enabling downstream pairwise computations on the coarse grid ($O(L^2/s^2)$). This design reduces memory and computational cost without compromising generative accuracy. Our approach achieves state-of-the-art performance on CrossDocked2020 and DUD-E benchmarks while significantly reducing pairwise-tensor memory. SiDGen bridges the gap between sequence-based efficiency and pocket-aware conditioning, offering a scalable path for high-throughput structure-based discovery.