loading page

zPoseScore model for accurate and robust protein-ligand docking pose scoring in CASP15
  • +10
  • Liangzhen Zheng,
  • Tao Shen,
  • Fuxu Liu,
  • Zechen Wang,
  • Jinyuan Sun,
  • Yifan Bu,
  • Jintao Meng,
  • Weihua Chen,
  • Yuguang Mu,
  • Weifeng Li,
  • Guoping Zhao,
  • Sheng Wang,
  • Wei Yanjie
Liangzhen Zheng
Shanghai Zelixir Biotech Company Ltd

Corresponding Author:[email protected]

Author Profile
Tao Shen
Shanghai Zelixir Biotech Company Ltd
Author Profile
Fuxu Liu
Shanghai Zelixir Biotech Company Ltd
Author Profile
Zechen Wang
Shandong University
Author Profile
Jinyuan Sun
Institute of Microbiology Chinese Academy of Sciences
Author Profile
Yifan Bu
Shanghai Zelixir Biotech Company Ltd
Author Profile
Jintao Meng
Chinese Academy of Sciences Shenzhen Institutes of Advanced Technology
Author Profile
Weihua Chen
Shanghai Zelixir Biotech Company Ltd
Author Profile
Yuguang Mu
Nanyang Technological University School of Biological Sciences
Author Profile
Weifeng Li
Shandong University
Author Profile
Guoping Zhao
Chinese Academy of Sciences Shenzhen Institutes of Advanced Technology
Author Profile
Sheng Wang
Author Profile
Wei Yanjie
Chinese Academy of Sciences Shenzhen Institutes of Advanced Technology
Author Profile

Abstract

We introduce a deep learning-based ligand pose scoring model called zPoseScore for predicting protein-ligand complexes in the 15th Critical Assessment of Protein Structure Prediction (CASP15). Our contributions are three-fold: firstly, we generate six training and evaluation datasets by employing advanced data augmentation and sampling methods. Secondly, we redesign the “zFormer” module, inspired by AlphaFold2’s Evoformer, to efficiently describe protein-ligand interactions. This module enables the extraction of protein-ligand paired features that lead to accurate predictions. Lastly, we develop the zPoseScore framework with zFormer for scoring and ranking ligand poses, allowing for atomic-level protein-ligand feature encoding and fusion to output refined ligand poses and ligand per-atom deviations. Our results demonstrate excellent performance on various testing datasets, achieving Pearson’s correlation R = 0.783 and 0.659 for ranking docking decoys generated based on experimental and predicted protein structures of CASF-2016 protein-ligand complexes. Additionally, we obtain an averaged lDDT = 0.558 of AIchemy_LIG2 in CASP15 for de novo protein-ligand complex structure predictions. Detailed analysis shows that accurate ligand binding site prediction and side-chain orientation are crucial for achieving better prediction performance. Our proposed model is one of the most accurate protein-ligand pose prediction models and could serve as a valuable tool in small molecule drug discovery.
16 Apr 2023Submitted to PROTEINS: Structure, Function, and Bioinformatics
17 Apr 2023Submission Checks Completed
17 Apr 2023Assigned to Editor
17 Apr 2023Review(s) Completed, Editorial Evaluation Pending
19 Apr 2023Reviewer(s) Assigned
09 May 2023Editorial Decision: Revise Minor
08 Jun 20231st Revision Received
12 Jun 2023Assigned to Editor
12 Jun 2023Submission Checks Completed
12 Jun 2023Review(s) Completed, Editorial Evaluation Pending
12 Jun 2023Reviewer(s) Assigned
26 Jun 2023Editorial Decision: Revise Minor
21 Jul 20232nd Revision Received
21 Jul 2023Submission Checks Completed
21 Jul 2023Assigned to Editor
21 Jul 2023Review(s) Completed, Editorial Evaluation Pending
24 Jul 2023Reviewer(s) Assigned
31 Jul 2023Editorial Decision: Accept