CoChat
Sign Up
PreprintCited by 106Computational Drug Discovery MethodsProtein Structure and DynamicsProtein Degradation and Inhibitors

The DOI matches its Crossref record; 2 of the 2 references it cites were found in a registry or an index.

SurfDock is a Surface-Informed Diffusion Generative Model for Reliable and Accurate Protein-ligand Complex Prediction

Open access
Cao, Duanhua, Chen, Mingan, Zhang, Runze, Yu, Jie, Jiang, Xinyu, Fan, Zhehuan, Zhang, Wei, Zheng, Mingyue·2023·Nature Methods
DOI 10.1101/2023.12.13.571408

Preprint — published as 10.1038/s41592-024-02516-y · Paper page

Open this paper in CoChat — free accountOpen PDF

Abstract

In the field of structure-based drug design, accurately predicting the binding conformation of ligands to proteins is a long-standing objective. Despite recent advances in deep learning yielding various methods for predicting protein-ligand complex structures, these AI-driven approaches frequently fall short of traditional docking methods in practice and often yield structures that lack physical and chemical plausibility. To overcome these limitations, we present SurfDock, an advanced geometric diffusion network, distinguished by its ability to integrate multiple protein representations including protein sequence, three-dimensional structural graphs, and surface-level details into its equivariant architecture. SurfDock employs a generative diffusion model on a non-Euclidean manifold, enabling precise optimization of molecular translations, rotations, and torsions for reliable binding poses generation. Complemented by a mixture density network for scoring using the same comprehensive representation, SurfDock achieves significantly improved docking success rates over all existing methods, excelling in both accuracy and adherence to physical constraints. Equipped with post-docking energy minimization as an optional feature, the plausibility of generated poses is further enhanced. Importantly, SurfDock demonstrates excellent generalizability to unseen proteins and extensibility to virtual screening tasks with state-of-the-art performance. We consider it a transformative contribution that could serve as an invaluable asset in structure-based drug design.

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 · 10.1101/2023.12.13.571408

Cites

2 of 2 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: 54

  1. 2023, De novo design of protein structure and function with RFdiffusion, Nature, 620, 1089, 10.1038/s41586-023-06415-8 · DOI · Paper page
  2. 10.1126/science.abj8754 · DOI · Paper page
  3. 10.1038/s41586-023-05905-z · DOI
  4. 2023, Structure-based drug design with geometric deep learning, Curr. Opin. Struct. Biol, 79, 102548, 10.1016/j.sbi.2023.102548 · DOI · Paper page
  5. 10.1038/384023a0 · DOI
  6. 10.1038/s41586-022-05258-z · DOI
  7. 2019, Ultra-large library docking for discovering new chemotypes, Nature, 566, 224, 10.1038/s41586-019-0917-9 · DOI · Paper page
  8. 2023, A practical guide to machine-learning scoring for structure-based virtual screening, Nat. Protoc, 18, 3460, 10.1038/s41596-023-00885-w · DOI
  9. 10.1021/acs.jcim.1c00203 · DOI
  10. 10.1021/jm0306430 · DOI · Paper page
  11. 10.1002/prot.10465 · DOI
  12. 2023, Deep learning model for efficient protein-ligand docking with implicit side-chain flexibility, J. Chem. Inf. Model, 63, 1695, 10.1021/acs.jcim.2c01436 · DOI
  13. 10.1021/ci300604z · DOI
  14. 10.1021/acs.jcim.6b00740 · DOI
  15. 2021, A geometric deep learning approach to predict binding conformations of bioactive molecules, Nat. Mach. Intell, 3, 1033, 10.1038/s42256-021-00409-9 · DOI
  16. Cao, D. et al. EquiScore: A generic protein-ligand interaction scoring method integrating physical prior knowledge with data augmentation modeling. Preprint at https://www.biorxiv.org/content/10.1101/2023.06.18.545464v2 (2023). · DOI
  17. Li, H. , Sze, K.-H. , Lu, G. & Ballester, P.J. Machine-learning scoring functions for structure-based virtual screening. Wiley Interdiscip. Rev. Comput. Mol. Sci. 11 (2021). · DOI
  18. 2021, Improving structure-based virtual screening with ensemble docking and machine learning, J. Chem. Inf. Model, 61, 5362, 10.1021/acs.jcim.1c00511 · DOI
  19. Xiong, G. et al. Featurization strategies for protein–ligand interactions and their applications in scoring function development. Wiley Interdiscip. Rev. Comput. Mol. Sci. 12 (2022). · DOI
  20. Zhou, G. et al. Uni-Mol: A Universal 3D Molecular Representation Learning Framework. In International Conference on Learning Representations (2023). · DOI
  21. Stärk, H. , Ganea, O. , Pattanaik, L. , Barzilay, R. & Jaakkola, T. Equibind: Geometric deep learning for drug binding structure prediction. In International Conference on Machine Learning, 20503–20521 (2022).
  22. Zhang, Y. , Cai, H. , Shi, C. & Tang, J. E3Bind: An End-to-End Equivariant Network for Protein-Ligand Docking. In International Conference on Learning Representations (2023).
  23. 2022, TANKBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction, In Advances in Neural Information Processing Systems, 35, 7236
  24. 2023, Efficient and accurate large library ligand docking with KarmaDock, Nat. Comput. Sci, 3, 789, 10.1038/s43588-023-00511-5 · DOI
  25. Corso, G. , Stärk, H. , Jing, B. , Barzilay, R. & Jaakkola, T.S. DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking. In The Eleventh International Conference on Learning Representations (2023).
  26. 2020, Denoising diffusion probabilistic models, In Advances in Neural Information Processing Systems, 33, 6840
  27. Guan, J. et al. 3D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Prediction. In The 11th International Conference on Learning Representations (2023).
  28. Xu, M. et al. GeoDiff: A Geometric Diffusion Model for Molecular Conformation Generation. In International Conference on Learning Representations (2022).
  29. Hoogeboom, E. , Satorras, V.G. , Vignac, C. & Welling, M. Equivariant diffusion for molecule generation in 3d. In International Conference on Machine Learning, 8867–8887 (2022).
  30. Yu, Y. , Lu, S. , Gao, Z. , Zheng, H. & Ke, G. Do deep learning models really outperform traditional approaches in molecular docking? In International Conference on Learning Representations (2023).
  31. 10.1093/bioinformatics/btv082 · DOI
  32. Buttenschoen, M. , Morris, G.M. & Deane, C.M. PoseBusters: AI-based docking methods fail to generate physically valid poses or generalise to novel sequences. Preprint at https://arxiv.org/abs/2308.05777 (2023). · DOI
  33. Harris, C. et al. PoseCheck: Generative Models for 3D Structure-based Drug Design Produce Unrealistic Poses. In NeurIPS 2023 Generative AI and Biology (GenBio) Workshop (2023).
  34. 2021, Geometric deep learning on molecular representations, Nat. Mach. Intell, 3, 1023, 10.1038/s42256-021-00418-8 · DOI
  35. 2020, Deciphering interaction fingerprints from protein molecular surfaces using geometric deep learning, Nat. Methods, 17, 184, 10.1038/s41592-019-0666-6 · DOI
  36. 2022, ScanNet: an interpretable geometric deep learning model for structure-based protein binding site prediction, Nat. Methods, 19, 730, 10.1038/s41592-022-01490-7 · DOI · Paper page
  37. 10.1038/s41586-023-05993-x · DOI
  38. 2023, Learning on topological surface and geometric structure for 3D molecular generation, Nat. Comput. Sci, 3, 849, 10.1038/s43588-023-00530-2 · DOI
  39. Sverrisson, F. , Feydy, J. , Correia, B.E. & Bronstein, M.M. Fast end-to-end learning on protein surfaces. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 15267–15276 (2021). · DOI
  40. 10.1093/bioinformatics/btaa524 · DOI · Paper page
  41. 2023, Sequence-based drug design as a concept in computational drug design, Nat. Commun, 14, 4217, 10.1038/s41467-023-39856-w · DOI
  42. Bryant, P. , Kelkar, A. , Guljas, A. , Clementi, C. & Noé, F. Structure prediction of protein-ligand complexes from sequence information with Umol. Preprint at https://www.biorxiv.org/content/10.1101/2023.11.03.565471v1 (2023). · DOI
  43. 10.1021/jm030580l · DOI
  44. 10.1021/jm061277y · DOI
  45. 10.1021/ci400115b · DOI · Paper page
  46. 10.1126/science.ade2574 · DOI · Paper page
  47. 2020, Tapping on the black box: How is the scoring power of a machine-learning scoring function dependent on the training set?, J. Chem. Inf. Model, 60, 1122, 10.1021/acs.jcim.9b00714 · DOI
  48. 10.1039/c6cp01555g · DOI · Paper page
  49. 2022, E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials, Nat. Commun, 13, 2453, 10.1038/s41467-022-29939-5 · DOI · Paper page
  50. 2022, Torsional Diffusion for Molecular Conformer Generation, In Advances in Neural Information Processing Systems, 35, 24240
  51. Geiger, M. & Smidt, T. e3nn: Euclidean neural networks. Preprint at https://arxiv.org/abs/2207.09453 (2022). · arXiv
  52. 10.1371/journal.pcbi.1005659 · DOI · Paper page
  53. 2023, Development and benchmarking of Open Force Field 2.0.0: The Sage small molecule force field, J. Chem. Theory Comput, 19, 3251, 10.1021/acs.jctc.3c00039 · DOI
  54. 10.1002/jcc.20035 · DOI

Metadata

Authors
Cao, D., Chen, M., Zhang, R., Yu, J., Jiang, X., Fan, Z., Zhang, W., & Zheng, M.
Affiliations
Chinese Academy of Sciences (CN)
Corresponding author
Zheng, Mingyue
ORCID
Venue
Nature Methods
Publisher
openRxiv
Published
2023-12-14
Volume
22
Issue
2
Pages
310-322
ISSN
1548-7091 (ISSN-L)
DOI
10.1101/2023.12.13.571408
OpenAlex
W4404764874
Cited by
106
References
2
Citations by year
2026: 60 · 2025: 44 · 2024: 2
FWCI
22.47
Type
journal article
Language
en
Open access
Licence
cc-by-nd
Locations
  • Nature Methods (publishedVersion)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nd)
  • PubMed (publishedVersion)
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
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Degradation and Inhibitors
Keywords
protein-ligand docking, generative diffusion models, equivariant neural networks, non-Euclidean manifolds, binding pose prediction, virtual screening, apo structure
MeSH terms
Deep Learning, Algorithms, Humans, Ligands, Protein Binding, Protein Conformation, Proteins, Software, Molecular Docking 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