Disentangled diffusion model for 3D molecular generation with protein–ligand interaction priors
Bioinformatics · 01.06.2026 · 42 (6)
DOI: 10.1093/bioinformatics/btag165 ↗Kurz gefasst
Fachartikel in der Zeitschrift Bioinformatics, online erschienen am 05.05.2026. Verfasst von Zhilin Huang und 9 weiteren Autor:innen. Beteiligte Einrichtung: Shenzhen International Graduate School, Tsinghua University, Shenzhen, 518000, China. Hauptthemen laut MEDLINE: Proteins, Models, Molecular, Drug Design, Computational Biology. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Motivation Structure-based drug design (SBDD) aims to generate ligand molecules that tightly bind to specific protein targets, a critical step in drug discovery. Diffusion models have shown promise for this task, yet existing methods struggle to effectively incorporate protein–ligand interaction priors during generation. Most approaches rely on protein-specific structural priors that remain fixed throughout generation, limiting molecular diversity and failing to capture the dynamic interplay between protein pockets and ligand atoms, which is essential for achieving high binding affinity. Results We propose DPDiff, a disentangled prior-conditioned diffusion model for protein-specific 3D molecular generation. DPDiff introduces two complementary interaction prior networks that capture geometry-based spatial interactions and sequence-based interactions robust to structural noise. During generation, the model dynamically extracts interaction priors using intermediate diffusion predictions and adaptively fuses them via a time-dependent adapter. A disentangled denoising network balances prior guidance with generative flexibility. Experiments on the CrossDocked2020 dataset demonstrate that DPDiff generates molecules with more realistic 3D structures and state-of-the-art binding affinities, achieving an average Vina Dock score of −8.58 and a high affinity ratio of 69.4%, outperforming existing methods while maintaining favorable drug-likeness and synthetic accessibility. Availability and implementation The source code of DPDiff is available at https://github.com/ZerinHwang03/DPDiff.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Shenzhen International Graduate School, Tsinghua University, Shenzhen, 518000, China.
- Institute of Perceptual Intelligence, Peng Cheng Laboratory, Shenzhen, 518000, China.
- Institute of Medical Technology, Peking University, Beijing, 100080, China.
- China Southern Power Dispatching and Control Center, China Southern Power Grid, Guangzhou, 510700, China.
- Institute of Computing, University of Chinese Academy of Sciences, Beijing, 100080, China.
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, 100080, China.
ORCID-Profile: Zhilin Huang, Xin Gao, Wenming Yang
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Huang, Z., Yang, L., Qin, C., Xing, Y., Yu, H., Zhou, X., Zheng, B., Wang, Y., Gao, X., & Yang, W. (2026). Disentangled diffusion model for 3D molecular generation with protein–ligand interaction priors. Bioinformatics, 42(6). https://doi.org/10.1093/bioinformatics/btag165
Vancouver
Huang Z, Yang L, Qin C, Xing Y, Yu H, Zhou X, et al. Disentangled diffusion model for 3D molecular generation with protein–ligand interaction priors. Bioinformatics. 2026;42(6). doi:10.1093/bioinformatics/btag165
BibTeX
@article{Huang2026,
author = {Huang, Zhilin and Yang, Ling and Qin, Chujun and Xing, Yifei and Yu, Hongyang and Zhou, Xiangxin and Zheng, Bing and Wang, Yu and Gao, Xin and Yang, Wenming},
title = {Disentangled diffusion model for 3D molecular generation with protein–ligand interaction priors},
journal = {Bioinformatics},
year = {2026},
volume = {42},
number = {6},
doi = {10.1093/bioinformatics/btag165}
}
EUVIMED