CoChat
Sign Up
Cited by 6Computational Drug Discovery MethodsProtein Structure and DynamicsModel Reduction and Neural Networks

The DOI matches its Crossref record; 3 of the 3 references it cites were found in a registry or an index.

A multi-grained symmetric differential equation model for learning protein-ligand binding dynamics

Open access
Liu, Shengchao, Du, Weitao, Xu, Hannan, Li, Yanjing, Li, Zhuoxinran, Bhethanabotla, Vignesh, Yan, Divin, Borgs, Christian, Anandkumar, Anima, Guo, Hongyu, Chayes, Jennifer·2025·Nature Communications
DOI 10.1038/s41467-025-67808-zPMC PMC12848303

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Molecular dynamics (MD) simulation is a key tool in drug discovery for predicting protein-ligand binding affinities, transport properties, and pocket dynamics. While advances in numerical and machine learning (ML) methods have improved MD efficiency, accurately modeling long-timescale dynamics remains challenging. We introduce NeuralMD, an ML surrogate that accelerates and enhances MD simulations of protein-ligand binding. NeuralMD employs a physics-informed, multi-grained, group-symmetric framework comprising (1) BindingNet, which enforces symmetry via vector frames and captures multi-level protein-ligand interactions, and (2) an augmented neural differential equation solver that learns trajectories under Newtonian mechanics. Across ten single-trajectory and three multi-trajectory tasks, NeuralMD achieves up to 15 × lower reconstruction error and 70% higher validity than existing ML baselines. The predicted oscillations closely align with ground-truth dynamics, establishing NeuralMD as a foundation for next-generation protein-ligand simulation research.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-025-67808-z

Cites

3 of 3 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 48

  1. Kairys, V., Baranauskiene, L., Kazlauskiene, M., Matulis, D. & Kazlauskas, E. Binding affinity in drug design: experimental and computational techniques. Expert Opin. drug Discov. 14, 755–768 (2019). · DOI
  2. Lahey, S.-L. J. & Rowley, C. N. Simulating protein–ligand binding with neural network potentials. Chem. Sci. 11, 2362–2368 (2020). · DOI
  3. Yang, J., Shen, C. & Huang, N. Predicting or pretending: artificial intelligence for protein-ligand interactions lack of sufficiently large and unbiased datasets. Front. Pharmacol. 11, 69 (2020). · DOI
  4. Volkov, M. et al. On the frustration to predict binding affinities from protein–ligand structures with deep neural networks. J. Med. Chem. 65, 7946–7958 (2022). · DOI
  5. Lifson, S. & Warshel, A. Consistent force field for calculations of conformations, vibrational spectra, and enthalpies of cycloalkane and n-alkane molecules. J. Chem. Phys. 49, 5116–5129 (1968). · DOI
  6. Kohn, W. Nobel lecture: Electronic structure of matter-wave functions and density functionals. Rev. Mod. Phys. 71, 1253 (1999). · DOI
  7. Behler, J. & Parrinello, M. Generalized neural-network representation of high-dimensional potential-energy surfaces. Phys. Rev. Lett. 98, 146401 (2007). · DOI · Paper page
  8. Chmiela, S. et al. Machine learning of accurate energy-conserving molecular force fields. Sci. Adv. 3, e1603015 (2017). · DOI
  9. Zhang, L., Han, J., Wang, H., Car, R. & Weinan, E. Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics. Phys. Rev. Lett. 120, 143001 (2018). · DOI
  10. Liu, S., Yan, D., Guo, H. & Anandkumar, A. An equivariant flow matching framework for learning molecular crystallization. In ICML 2024 Workshop on Geometry-grounded Representation Learning and Generative Modeling (ICML, 2024).
  11. Siebenmorgen, T. et al. Misato: machine learning dataset of protein–ligand complexes for structure-based drug discovery. Nature Computational Science 4, 367–378 (2024). · DOI · Paper page
  12. Schütt, K. T., Sauceda, H. E., Kindermans, P.-J., Tkatchenko, A. & Müller, K.-R. Schnet–a deep learning architecture for molecules and materials. J. Chem. Phys. 148, 241722 (2018). · DOI · Paper page
  13. Smidt, T. et al. Tensor field networks: rotation-and translation-equivariant neural networks for 3D point clouds. arXiv preprint arXiv:1802.08219 (2018). · arXiv
  14. Klicpera, J., Giri, S., Margraf, J. T. & Günnemann, S. Fast and uncertainty-aware directional message passing for non-equilibrium molecules. arXiv preprint arXiv:2011.14115 (2020). · arXiv
  15. Qiao, Z., Welborn, M., Anandkumar, A., Manby, F. R. & Miller, T. F. Orbnet: deep learning for quantum chemistry using symmetry-adapted atomic-orbital features. J. Chem. Phys. 153, 124111 (2020). · DOI
  16. Musaelian, A. et al. Learning local equivariant representations for large-scale atomistic dynamics. Nat. Commun. 14, 579 (2023). · DOI
  17. Liu, S. et al. Symmetry-informed geometric representation for molecules, proteins, and crystalline materials. Adv. Neural Inf. Process. Syst. 36, 66084–66101 (2023).
  18. Liu, S., Du, W., Ma, Z.-M., Guo, H. & Tang, J. A group symmetric stochastic differential equation model for molecule multi-modal pretraining. In International Conference on Machine Learning, 21497–21526 (PMLR, 2023).
  19. Du, W., Chen, J., Zhang, X., Ma, Z.-M. & Liu, S. Molecule joint auto-encoding: trajectory pretraining with 2d and 3d diffusion. Adv. Neural Inf. Process. Syst. 36, 55077–55096 (2023).
  20. Vajda, S., Beglov, D., Wakefield, A. E., Egbert, M. & Whitty, A. Cryptic binding sites on proteins: definition, detection, and druggability. Curr. Opin. Chem. Biol. 44, 1–8 (2018). · DOI
  21. Fu, X. et al. Forces are not enough: benchmark and critical evaluation for machine learning force fields with molecular simulations. Transactions on Machine Learning Research (NIPS, 2023).
  22. Chen, R. T., Rubanova, Y., Bettencourt, J. & Duvenaud, D. K. Neural ordinary differential equations. Adv. Neural Inf. Process. Syst. 31, 6572–6583 (2018).
  23. Raissi, M., Perdikaris, P. & Karniadakis, G. E. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. 378, 686–707 (2019). · DOI · Paper page
  24. Lu, L., Jin, P., Pang, G., Zhang, Z. & Karniadakis, G. E. Learning nonlinear operators via deeponet based on the universal approximation theorem of operators. Nat. Mach. Intell. 3, 218–229 (2021). · DOI
  25. Li, Z. et al. Neural operator: graph kernel network for partial differential equations. arXiv preprint arXiv:2003.03485 (2020). · arXiv
  26. Kovachki, N. et al. Neural operator: learning maps between function spaces. J. Mach. Learn. Res. 24, 1–97 (2023).
  27. Liu, S. et al. Manifold-constrained nucleus-level denoising diffusion model for structure-based drug design. Proc. Natl. Acad. Sci. USA 122, e2415666122 (2025). · DOI
  28. Salmaso, V. & Moro, S. Bridging molecular docking to molecular dynamics in exploring ligand-protein recognition process: an overview. Front. Pharmacol. 9, 923 (2018). · DOI
  29. Berman, H. M. et al. The protein data bank. Nucleic Acids Res. 28, 235–242 (2000). · DOI · Paper page
  30. Musaelian, A., Johansson, A., Batzner, S. & Kozinsky, B. Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 1–12 (SC, 2023). · DOI
  31. Fu, X., Xie, T., Rebello, N. J., Olsen, B. D. & Jaakkola, T. Simulate time-integrated coarse-grained molecular dynamics with geometric machine learning. In ICLR Workshop on Deep Generative Models for Highly Structured Data (ICLR, 2022).
  32. Wu, F. & Li, S. Z. DiffMD: A geometric diffusion model for molecular dynamics simulations. In Proceedings of the AAAI Conference on Artificial Intelligence, 5321–5329 (AAAI, 2023). · DOI
  33. Arts, M. et al. Two for one: Diffusion models and force fields for coarse-grained molecular dynamics. J. Chem. Theory Comput. 19, 6151–6159 (2023). · DOI
  34. Van Der Spoel, D. et al. Gromacs: fast, flexible, and free. J. Comput. Chem. 26, 1701–1718 (2005). · DOI
  35. Liu, T., Zhang, L., Joo, D. & Sun, S.-C. Nf-κb signaling in inflammation. Signal Transduct. Target. Ther. 2, 1–9 (2017).
  36. Mulero, M. C., Wang, V. Y.-F., Huxford, T. & Ghosh, G. Genome reading by the nf-κb transcription factors. Nucleic Acids Res. 47, 9967–9989 (2019). · DOI
  37. de Leon-Boenig, G. et al. The crystal structure of the catalytic domain of the nf-κb inducing kinase reveals a narrow but flexible active site. Structure 20, 1704–1714 (2012). · DOI
  38. Park, M. H. & Hong, J. T. Roles of nf-κb in cancer and inflammatory diseases and their therapeutic approaches. Cells 5, 15 (2016). · DOI
  39. Hayden, M. S. & Ghosh, S. Nf-κb in immunobiology. Cell Res. 21, 223–244 (2011). · DOI
  40. Thu, Y. M. & Richmond, A. Nf-κb inducing kinase: a key regulator in the immune system and in cancer. Cytokine Growth Factor Rev. 21, 213–226 (2010). · DOI
  41. Haselager, M. V. & Eldering, E. The therapeutic potential of targeting nik in b cell malignancies. Front. Immunol. 13, 930986 (2022). · DOI
  42. Crawford, J. J. et al. Filling a nick in nik: Extending the half-life of a nik inhibitor through structure-based drug design. Bioorg. Med. Chem. Lett. 89, 129277 (2023). · DOI
  43. Pflug, K. M. & Sitcheran, R. Targeting nf-κb-inducing kinase (nik) in immunity, inflammation, and cancer. Int. J. Mol. Sci. 21, 8470 (2020). · DOI
  44. Cheng, J. et al. Pharmacological inhibition of nf-κb-inducing kinase (nik) with small molecules for the treatment of human diseases. RSC Med. Chem. 12, 552–565 (2021). · DOI
  45. Li, K. et al. Inhibiting nf-κb-inducing kinase (nik): discovery, structure-based design, synthesis, structure–activity relationship, and co-crystal structures. Bioorg. Med. Chem. Lett. 23, 1238–1244 (2013). · DOI
  46. Hsu, E. P.Stochastic Analysis on Manifolds. 38 (American Mathematical Society, 2002). · DOI
  47. Unke, O. T. et al. Machine learning force fields. Chem. Rev. 121, 10142–10186 (2021). · DOI · Paper page
  48. Stocker, S., Gasteiger, J., Becker, F., Günnemann, S. & Margraf, J. T. How robust are modern graph neural network potentials in long and hot molecular dynamics simulations? Mach. Learn. Sci. Technol. 3, 045010 (2022). · DOI

Metadata

Authors
Liu, S., Du, W., Xu, H., Li, Y., Li, Z., Bhethanabotla, V., Yan, D., Borgs, C., Anandkumar, A., Guo, H., & Chayes, J.
Affiliations
  • University of California, Berkeley (US)
  • University of Oxford (GB)
  • University of Toronto (CA)
  • California Institute of Technology (US)
  • National Research Council Canada (CA)
Corresponding author
  • Liu, Shengchao
  • Borgs, Christian
  • Guo, Hongyu
  • Chayes, Jennifer
ORCID
Venue
Nature Communications
Publisher
Springer Science and Business Media LLC
Published
2025-12-30
Volume
17
Issue
1
Pages
1049-1049
ISSN
2041-1723 (ISSN-L)
DOI
10.1038/s41467-025-67808-z
PMC
PMC12848303
OpenAlex
W7117551302
Cited by
6
References
3
Citations by year
2026: 3 · 2025: 3
FWCI
3.09
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • Research Square (acceptedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Life Sciences › Pharmacology, Toxicology and Pharmaceutics › Drug Discovery › Computational Drug Discovery Methods
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Model Reduction and Neural Networks
Keywords
neural differential equations, physics-informed machine learning, group symmetry, molecular dynamics acceleration, binding affinity prediction, long-timescale molecular dynamics
MeSH terms
Machine Learning, Ligands, Protein Binding, Proteins, Neural Networks, Computer, Drug Discovery, Molecular Dynamics Simulation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review