Predicting the protein-ligand affinity from molecular dynamics trajectories

Authors:
  • YMYaosen Min
    Yaosen Min
  • YWYe Wei
    Ye Wei
  • PWPeizhuo Wang
    Peizhuo Wang
  • NWNian Wu
    Nian Wu
  • SBStefan Bauer
    Stefan Bauer
  • SZShuxin Zheng
    Shuxin Zheng
  • YSYu Shi
    Yu Shi
  • YWYingheng Wang
    Yingheng Wang
  • DZDan Zhao
    Dan Zhao
  • JWJi Wu
    Ji Wu
  • JZJianyang Zeng
    Jianyang Zeng

Abstract
The accurate protein-ligand binding affinity prediction is essential in drug design and many other molecular recognition problems. Despite many advances in affinity prediction based on machine learning techniques, they are still limited since the protein-ligand binding is determined by the dynamics of atoms and molecules. To this end, we curated an MD dataset containing 3,218 dynamic protein-ligand complexes and further developed Dynaformer, a graph-based deep learning framework. Dynaformer can fully capture the dynamic binding rules by considering various geometric characteristics of the interaction. Our method shows superior performance over the methods hitherto reported. Moreover, we performed virtual screening on heat shock protein 90 (HSP90) by integrating our model with structure-based docking. We benchmarked our performance against other baselines, demonstrating that our method can identify the molecule with the highest experimental potency. We anticipate that large-scale MD dataset and machine learning models will form a new synergy, providing a new route towards accelerated drug discovery and optimization.