DecompDiff: Diffusion Models with Decomposed Priors for Structure-Based Drug Design

Jiaqi Guan, Xiangxin Zhou, Yuwei Yang, Yu Bao, Jian Peng, Jianzhu Ma, Qiang Liu, Liang Wang, Quanquan Gu
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:11827-11846, 2023.

Abstract

Designing 3D ligands within a target binding site is a fundamental task in drug discovery. Existing structured-based drug design methods treat all ligand atoms equally, which ignores different roles of atoms in the ligand for drug design and can be less efficient for exploring the large drug-like molecule space. In this paper, inspired by the convention in pharmaceutical practice, we decompose the ligand molecule into two parts, namely arms and scaffold, and propose a new diffusion model, DecompDiff, with decomposed priors over arms and scaffold. In order to facilitate the decomposed generation and improve the properties of the generated molecules, we incorporate both bond diffusion in the model and additional validity guidance in the sampling phase. Extensive experiments on CrossDocked2020 show that our approach achieves state-of-the-art performance in generating high-affinity molecules while maintaining proper molecular properties and conformational stability, with up to $-8.39$ Avg. Vina Dock score and $24.5%$ Success Rate. The code is provided at https://github.com/bytedance/DecompDiff

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-guan23a, title = {{D}ecomp{D}iff: Diffusion Models with Decomposed Priors for Structure-Based Drug Design}, author = {Guan, Jiaqi and Zhou, Xiangxin and Yang, Yuwei and Bao, Yu and Peng, Jian and Ma, Jianzhu and Liu, Qiang and Wang, Liang and Gu, Quanquan}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {11827--11846}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/guan23a/guan23a.pdf}, url = {https://proceedings.mlr.press/v202/guan23a.html}, abstract = {Designing 3D ligands within a target binding site is a fundamental task in drug discovery. Existing structured-based drug design methods treat all ligand atoms equally, which ignores different roles of atoms in the ligand for drug design and can be less efficient for exploring the large drug-like molecule space. In this paper, inspired by the convention in pharmaceutical practice, we decompose the ligand molecule into two parts, namely arms and scaffold, and propose a new diffusion model, DecompDiff, with decomposed priors over arms and scaffold. In order to facilitate the decomposed generation and improve the properties of the generated molecules, we incorporate both bond diffusion in the model and additional validity guidance in the sampling phase. Extensive experiments on CrossDocked2020 show that our approach achieves state-of-the-art performance in generating high-affinity molecules while maintaining proper molecular properties and conformational stability, with up to $-8.39$ Avg. Vina Dock score and $24.5%$ Success Rate. The code is provided at https://github.com/bytedance/DecompDiff} }
Endnote
%0 Conference Paper %T DecompDiff: Diffusion Models with Decomposed Priors for Structure-Based Drug Design %A Jiaqi Guan %A Xiangxin Zhou %A Yuwei Yang %A Yu Bao %A Jian Peng %A Jianzhu Ma %A Qiang Liu %A Liang Wang %A Quanquan Gu %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-guan23a %I PMLR %P 11827--11846 %U https://proceedings.mlr.press/v202/guan23a.html %V 202 %X Designing 3D ligands within a target binding site is a fundamental task in drug discovery. Existing structured-based drug design methods treat all ligand atoms equally, which ignores different roles of atoms in the ligand for drug design and can be less efficient for exploring the large drug-like molecule space. In this paper, inspired by the convention in pharmaceutical practice, we decompose the ligand molecule into two parts, namely arms and scaffold, and propose a new diffusion model, DecompDiff, with decomposed priors over arms and scaffold. In order to facilitate the decomposed generation and improve the properties of the generated molecules, we incorporate both bond diffusion in the model and additional validity guidance in the sampling phase. Extensive experiments on CrossDocked2020 show that our approach achieves state-of-the-art performance in generating high-affinity molecules while maintaining proper molecular properties and conformational stability, with up to $-8.39$ Avg. Vina Dock score and $24.5%$ Success Rate. The code is provided at https://github.com/bytedance/DecompDiff
APA
Guan, J., Zhou, X., Yang, Y., Bao, Y., Peng, J., Ma, J., Liu, Q., Wang, L. & Gu, Q.. (2023). DecompDiff: Diffusion Models with Decomposed Priors for Structure-Based Drug Design. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:11827-11846 Available from https://proceedings.mlr.press/v202/guan23a.html.

Related Material