- Physical Review Journals
All JournalsPhysics Magazine

PRX Life

Reuse & Permissions

It is not necessary to obtain permission to reuse this article or its components as it is available under the terms of the Creative Commons Attribution 4.0 International license. This license permits unrestricted use, distribution, and reproduction in any medium, provided attribution to the author(s) and the published article's title, journal citation, and DOI are maintained. Please note that some figures may have been included with permission from other third parties. It is your responsibility to obtain the proper permission from the rights holder directly for these figures.

Export citation

Export citation

Choose format for download:

Download Citation
  • Open Access

Adapting Co-Folding Models for Structure-Based Protein-Protein Docking Through Flow Matching

Da Xu1, Lee-Shin Chu1, and Jeffrey J. Gray1,2,*

  • *Contact author: jgray@jhu.edu
PDF

PRX Life 4, 033018 – Published 24 August, 2026

DOI: https://doi.org/10.1103/zfn9-18y7

Abstract

Co-folding models like AlphaFold have revolutionized protein complex structure prediction, yet their reliance on multiple sequence alignments (MSAs) limits their applicability on challenging targets such as antibody-antigen complexes. An alternative approach, structure-based protein-protein docking, predicts the bound complex structure from the unbound monomer structures without requiring MSAs. In this work, we propose a method to adapt co-folding models for structure-based protein-protein docking by replacing their template module with a docking module, followed by training end-to-end with a flow-matching objective. We apply our method to AlphaFold-Multimer (AF-M) using the OpenFold implementation and transform it into a generative docking model, which we name AF2Dock. We evaluate AF2Dock and various baseline methods on the PINDER-AF2 benchmark and an antibody/nanobody test set. When using non-holo inputs, AF2Dock shows competitive performance compared to other structure-based docking methods and, in the case of nanobody complexes, outperforms all other docking methods tested here. Although AF2Dock underperforms co-folding AF-M and AF3 in success rates when using non-holo inputs, it produces orthogonal predictions and successfully identifies correct structures for targets where co-folding models fail. Ablation studies confirm that full-parameter fine-tuning of the AF-M components is critical for performance and reveal that, surprisingly, the inclusion of ESM embeddings can hinder success rates in certain cases such as nanobody complexes.

View figure in article

Physics Subject Headings (PhySH)

Article Text

Supplemental Material

References (42)

  1. R. Evans, M. O'Neill, A. Pritzel, N. Antropova, A. Senior, T. Green, A. Žídek, R. Bates, S. Blackwell, J. Yim, O. Ronneberger, S. Bodenstein, M. Zielinski, A. Bridgland, A. Potapenko, A. Cowie, K. Tunyasuvunakool, R. Jain, E. Clancy, P. Kohli, et al., Protein complex prediction with AlphaFold-multimer, bioRxiv (2021).
  2. J. Abramson, J. Adler, J. Dunger, R. Evans, T. Green, A. Pritzel, O. Ronneberger, L. Willmore, A. J. Ballard, J. Bambrick, S. W. Bodenstein, D. A. Evans, C.-C. Hung, M. O'Neill, D. Reiman, K. Tunyasuvunakool, Z. Wu, A. Žemgulytė, E. Arvaniti, C. Beattie, et al., Accurate structure prediction of biomolecular interactions with AlphaFold 3, Nature (London) 630, 493 (2024).
  3. C. Discovery, J. Boitreaud, J. Dent, M. McPartlon, J. Meier, V. Reis, A. Rogozhnikov, and K. Wu, Chai-1: Decoding the molecular interactions of life, bioRxiv (2024).
  4. S. Passaro, G. Corso, J. Wohlwend, M. Reveiz, S. Thaler, V. R. Somnath, N. Getz, T. Portnoi, J. Roy, H. Stark, D. Kwabi-Addo, D. Beaini, T. Jaakkola, and R. Barzilay, Boltz-2: Towards accurate and efficient binding affinity prediction, bioRxiv (2025).
  5. X. Chen, Y. Zhang, C. Lu, W. Ma, J. Guan, C. Gong, J. Yang, H. Zhang, K. Zhang, S. Wu, K. Zhou, Y. Yang, Z. Liu, L. Wang, B. Shi, S. Shi, and W. Xiao, Protenix-advancing structure prediction through a comprehensive AlphaFold3 reproduction, bioRxiv (2025).
  6. N. Corley, S. Mathis, R. Krishna, M. S. Bauer, T. R. Thompson, W. Ahern, M. W. Kazman, R. I. Brent, K. Didi, A. Kubaney, L. McHugh, A. Nagle, A. Favor, M. Kshirsagar, P. Sturmfels, Y. Li, J. Butcher, B. Qiang, L. L. Schaaf, R. Mitra, et al., Accelerating biomolecular modeling with AtomWorks and RF3, bioRxiv (2025).
  7. M. L. Rennie and M. R. Oliver, Emerging frontiers in protein structure prediction following the AlphaFold revolution, J. R. Soc. Interface 22, 20240886 (2025).
  8. F. Costa, M. Blum, and A. Bateman, Keeping it in the family: Using protein family templates to rescue low confidence AlphaFold2 models, Bioinform. Adv. 4, vbae188 (2024).
  9. R. Yin, B. Y. Feng, A. Varshney, and B. G. Pierce, Benchmarking AlphaFold for protein complex modeling reveals accuracy determinants, Protein Sci. 31, e4379 (2022).
  10. Z. Zhang, H. K. Wayment-Steele, G. Brixi, H. Wang, D. Kern, and S. Ovchinnikov, Protein language models learn evolutionary statistics of interacting sequence motifs, Proc. Natl. Acad. Sci. USA 121, e2406285121 (2024).
  11. I. A. Vakser, Protein-protein docking: From interaction to interactome, Biophys. J. 107, 1785 (2014).
  12. M. A. Ketata, C. Laue, R. Mammadov, H. Stärk, M. Wu, G. Corso, C. Marquet, R. Barzilay, and T. S. Jaakkola, DiffDock-PP: Rigid protein-protein docking with diffusion models arXiv:2304.03889.
  13. L.-S. Chu, S. Sarma, D. Xu, and J. J. Gray, Unified sampling and ranking for protein docking with DFMDock, bioRxiv (2024).
  14. M. McPartlon, C. Marquet, T. Geffner, D. Kovtun, A. Goncearenco, Z. Carpenter, L. Naef, M. Bronstein, and J. Xu, Latentdock: Protein-protein docking with latent diffusion, in Machine Learning in Structural Biology Workshop (NeurIPS, 2023).
  15. F. Sverrisson, M. Akdel, D. Abramson, J. Feydy, A. Goncearenco, Y. Adeshina, D. Kovtun, C. Marquet, X. Zhang, D. Baugher, Z. Carpenter, L. Naef, M. M. Bronstein, and B. Correia, Diffmasif: Surface-based protein-protein docking with diffusion models, in Machine Learning in Structural Biology Workshop (NeurIPS 2023).
  16. J. P. Roney and S. Ovchinnikov, State-of-the-art estimation of protein model accuracy using AlphaFold, Phys. Rev. Lett. 129, 238101 (2022).
  17. C. Mirabello, B. Wallner, B. Nystedt, S. Azinas, and M. Carroni, Unmasking AlphaFold to integrate experiments and predictions in multimeric complexes, Nat. Commun. 15, 8724 (2024).
  18. R. Yin and B. G. Pierce, Evaluation of AlphaFold antibody-antigen modeling with implications for improving predictive accuracy, Protein Sci. 33, e4865 (2024).
  19. Y. Lipman, R. T. Q. Chen, H. Ben-Hamu, M. Nickel, and M. Le, Flow matching for generative modeling, arXiv:2210.02747.
  20. G. Ahdritz, N. Bouatta, C. Floristean, S. Kadyan, Q. Xia, W. Gerecke, T. J. O'Donnell, D. Berenberg, I. Fisk, N. Zanichelli, B. Zhang, A. Nowaczynski, B. Wang, M. M. Stepniewska-Dziubinska, S. Zhang, A. Ojewole, M. E. Guney, S. Biderman, A. M. Watkins, S. Ra, et al., OpenFold: Retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization, Nat. Methods 21, 1514 (2024).
  21. D. Kovtun, M. Akdel, A. Goncearenco, G. Zhou, G. Holt, D. Baugher, D. Lin, Y. Adeshina, T. Castiglione, X. Wang, C. Marquet, M. McPartlon, T. Geffner, E. Rossi, G. Corso, H. Stärk, Z. Carpenter, E. Kucukbenli, M. Bronstein, and L. Naef, PINDER: The protein interaction dataset and evaluation resource, bioRxiv (2024).
  22. F. N. Hitawala and J. J. Gray, What does AlphaFold3 learn about antibody and nanobody docking, and what remains unsolved?, mAbs 17, 2545601 (2025).
  23. See Supplemental Material at http://link.aps.org/supplemental/10.1103/zfn9-18y7 for additional method details, figures, and algorithms.
  24. S. Candido, T. Hayes, A. Derry, R. Rao, Z. Lin, R. Verkuil, B. Z. Wu, J. S. Lee, E. S. Bruguera, J. A. Keval, M. Kopylov, J. E. Pak, W. Wu, N. Thomas, S. Mataraso, A. Hsu, A. C. Trotman-Grant, K. Fatras, A. dos Santos Costa, R. Badkundri, et al., Language modeling materializes a world model of protein biology, bioRxiv (2026).
  25. W. Peebles and S. Xie, Scalable diffusion models with transformers, arXiv:2212.09748.
  26. M. S. Albergo and E. Vanden-Eijnden, Building normalizing flows with stochastic interpolants, arXiv:2209.15571.
  27. X. Liu, C. Gong, and Q. Liu, Flow straight and fast: Learning to generate and transfer data with rectified flow, arXiv:2209.03003.
  28. B. Jing, B. Berger, and T. Jaakkola, AlphaFold meets flow matching for generating protein ensembles, arXiv:2402.04845.
  29. M. Mirdita, K. Schütze, Y. Moriwaki, L. Heo, S. Ovchinnikov, and M. Steinegger, ColabFold: Making protein folding accessible to all, Nat. Methods 19, 679 (2022).
  30. Y. Yan, H. Tao, J. He, and S.-Y. Huang, The HDOCK server for integrated protein-protein docking, Nat. Protoc. 15, 1829 (2020).
  31. B. G. Pierce, K. Wiehe, H. Hwang, B.-H. Kim, T. Vreven, and Z. Weng, ZDOCK server: Interactive docking prediction of protein-protein complexes and symmetric multimers, Bioinformatics 30, 1771 (2014).
  32. H. Hwang, T. Vreven, J. Janin, and Z. Weng, Protein-protein docking benchmark version 4.0, Proteins 78, 3111 (2010).
  33. R. F. Alford, A. Leaver-Fay, J. R. Jeliazkov, M. J. O'Meara, F. P. DiMaio, H. Park, M. V. Shapovalov, P. D. Renfrew, V. K. Mulligan, K. Kappel, J. W. Labonte, M. S. Pacella, R. Bonneau, P. Bradley, R. L. Dunbrack, Jr., R. Das, D. Baker, B. Kuhlman, T. Kortemme, and J. J. Gray, The Rosetta all-atom energy function for macromolecular modeling and design, J. Chem. Theory Comput. 13, 3031 (2017).
  34. V. Stebliankin, A. Shirali, P. Baral, J. Shi, P. Chapagain, K. Mathee, and G. Narasimhan, Evaluating protein binding interfaces with transformer networks, Nat. Mach. Intell. 5, 1042 (2023).
  35. D. Kuroda and J. J. Gray, Pushing the backbone in protein-protein docking, Structure 24, 1821 (2016).
  36. A. Harmalkar and J. J. Gray, Advances to tackle backbone flexibility in protein docking, Curr. Opin. Struct. Biol. 67, 178 (2021).
  37. D. Del Alamo, D. Sala, H. S. Mchaourab, and J. Meiler, Sampling alternative conformational states of transporters and receptors with AlphaFold2, eLife 11, e75751 (2022).
  38. H. K. Wayment-Steele, A. Ojoawo, R. Otten, J. M. Apitz, W. Pitsawong, M. Hömberger, S. Ovchinnikov, L. Colwell, and D. Kern, Predicting multiple conformations via sequence clustering and AlphaFold2, Nature (London) 625, 832 (2024).
  39. R. L. Dunbrack, Jr., Rēs ipSAE loquunt: What's wrong with AlphaFold's ipTM score and how to fix it, bioRxiv (2025).
  40. A. Harmalkar, S. Lyskov, and J. J. Gray, Reliable protein–protein docking with AlphaFold, rosetta, and replica exchange, eLife 13, RP94029 (2025).
  41. Code for AF2Dock is available at https://github.com/Graylab/AF2Dock.
  42. D. Xu, L.-S. Chu, and J. Gray, Model weights and input structures for “Adapting co-folding models for structure-based protein-protein docking through flow matching” [Figure], Zenodo, 2025, http://dx.doi.org/10.5281/ZENODO.17782958.

Outline

Information

PRX Life 4, 033018– Published 24 August, 2026
  • Received 5 January 2026
  • Accepted 8 July 2026
Crossmark - Check for updates button
Creative Commons Logo

Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI.

Published by the American Physical Society

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation