Unified Sampling and Ranking for Protein Docking With DFMDock
Proteins: Structure, Function, and Bioinformatics · 2026-12 · 94 (12) · 2013-2022
DOI: 10.1002/prot.70160 ↗At a glance
Journal article in Proteins: Structure, Function, and Bioinformatics, published online on 2026-08-06. Written by Lee-Shin Chu and 3 co-authors. Institution: Department of Chemical and Biomolecular Engineering, Johns Hopkins University, Baltimore, Maryland, USA. Major MEDLINE topics: Molecular Docking Simulation, Proteins, Computational Biology, Software. Cited 1 times so far (Crossref); 52 references. Funded by: NIGMS NIH HHS, Moderna, AstraZeneca.
Abstract (original language)
Recent diffusion‐based approaches to protein–protein docking typically decouple structure generation from decoy ranking. We introduce DFMDock (Denoising Force Matching for Docking), a unified diffusion model that integrates generative sampling and energy‐based ranking through physically motivated supervision. DFMDock predicts both denoising forces and a scalar energy, trained using force matching and energy contrastive objectives. The predicted forces guide the reverse diffusion process, while the energy enables decoy ranking without relying on a separately trained confidence model. On the Docking Benchmark 5, DFMDock achieves a 4.9% Top‐1 and 31.6% Oracle success rate, outperforming DiffDock‐PP (4.3% and 16.2%, respectively). Unlike co‐folding models, DFMDock does not require multiple sequence alignments (MSAs) and generalizes to unseen targets. In decoy ranking, its learned energy function outperforms Rosetta energy and model‐derived confidence scores, producing funnel‐shaped energy landscapes enriched for near‐native structures. These results suggest DFMDock as an efficient and physically grounded approach to diffusion‐based protein docking.
Subjects
MeSH terms (MEDLINE)
Substances
Authors and institutions
- Department of Chemical and Biomolecular Engineering, Johns Hopkins University, Baltimore, Maryland, USA.
- Program in Molecular Biophysics, Johns Hopkins University, Baltimore, Maryland, USA.
- Data Science and AI Institute, Johns Hopkins University, Baltimore, Maryland, USA.
- The Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, Maryland, USA.
Funding
- NIGMS NIH HHS
- Moderna
- AstraZeneca
- NIH HHS
- National Institutes of Health
Cite
APA
Chu, L. S., Sarma, S., Xu, D., & Gray, J. J. (2026). Unified Sampling and Ranking for Protein Docking With DFMDock. Proteins: Structure, Function, and Bioinformatics, 94(12), 2013-2022. https://doi.org/10.1002/prot.70160
Vancouver
Chu LS, Sarma S, Xu D, Gray JJ. Unified Sampling and Ranking for Protein Docking With DFMDock. Proteins: Structure, Function, and Bioinformatics. 2026;94(12):2013-2022. doi:10.1002/prot.70160
BibTeX
@article{Chu2026,
author = {Chu, Lee-Shin and Sarma, Sudeep and Xu, Da and Gray, Jeffrey J},
title = {Unified Sampling and Ranking for Protein Docking With DFMDock},
journal = {Proteins: Structure, Function, and Bioinformatics},
year = {2026},
volume = {94},
number = {12},
pages = {2013--2022},
doi = {10.1002/prot.70160}
}
EUVIMED