Multi-scale Iterative Refinement towards Robust and Versatile Molecular Docking

Jiaxian Yan,Zaixi Zhang,Kai Zhang,Qi Liu

2023-11-30

Abstract:Molecular docking is a key computational tool utilized to predict the binding conformations of small molecules to protein targets, which is fundamental in the design of novel drugs. Despite recent advancements in geometric deep learning-based approaches leading to improvements in blind docking efficiency, these methods have encountered notable challenges, such as limited generalization performance on unseen proteins, the inability to concurrently address the settings of blind docking and site-specific docking, and the frequent occurrence of physical implausibilities such as inter-molecular steric clash. In this study, we introduce DeltaDock, a robust and versatile framework designed for efficient molecular docking to overcome these challenges. DeltaDock operates in a two-step process: rapid initial complex structures sampling followed by multi-scale iterative refinement of the initial structures. In the initial stage, to sample accurate structures with high efficiency, we develop a ligand-dependent binding site prediction model founded on large protein models and graph neural networks. This model is then paired with GPU-accelerated sampling algorithms. The sampled structures are updated using a multi-scale iterative refinement module that captures both protein-ligand atom-atom interactions and residue-atom interactions in the following stage. Distinct from previous geometric deep learning methods that are conditioned on the blind docking setting, DeltaDock demonstrates superior performance in both blind docking and site-specific docking settings. Comprehensive experimental results reveal that DeltaDock consistently surpasses baseline methods in terms of docking accuracy. Furthermore, it displays remarkable generalization capabilities and proficiency for predicting physically valid structures, thereby attesting to its robustness and reliability in various scenarios.

Biomolecules,Machine Learning

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve

This paper aims to address several key challenges in molecular docking. Molecular docking is a crucial computational tool used to predict the binding conformations of small molecules with protein targets, playing a fundamental role in new drug design. Despite significant progress in the efficiency of blind docking using geometry-based deep learning methods, these methods still face the following major issues:

1. **Limited generalization performance on unseen proteins**: Existing methods perform poorly when dealing with new proteins that were not present in the training set.
2. **Inability to handle both blind docking and site-specific docking simultaneously**: Current geometric deep learning methods typically can only handle one docking scenario, either blind docking or docking with known binding sites.
3. **Physical unreasonableness**: Existing methods often generate unreasonable structures, such as steric clashes between molecules.

To overcome these issues, the authors propose DeltaDock, an efficient, robust, and versatile molecular docking framework. DeltaDock predicts the final binding structure through a two-step process: rapid initial complex structure sampling and multi-scale iterative refinement. Specifically:

- **Initial structure sampling**: By developing a ligand-based binding site prediction model (LigPoc), combined with large-scale pre-trained protein models and graph neural networks, it efficiently generates initial structures.
- **Multi-scale iterative refinement**: Through a multi-scale iterative refinement module, it captures protein-ligand atom-atom interactions and residue-atom interactions to further optimize the initial structure.

Experimental results show that DeltaDock performs excellently in both blind docking and site-specific docking, not only improving docking accuracy but also demonstrating outstanding generalization ability and the capability to generate physically reasonable structures. This makes DeltaDock highly promising for applications in drug discovery and design.