IAP-CFDock: Iterative Anchor Prediction and Coarse-to-Fine Protein-Ligand Blind Docking.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Protein-ligand blind docking is vital for drug discovery but challenging due to protein size and structural complexity.
• Accurate prediction of ligand binding poses is essential for identifying potential drug candidates.

Purpose of the Study:

• To introduce IAP-CFDock, a novel three-stage framework for accurate and efficient protein-ligand blind docking.
• To enhance anchor prediction accuracy using an efficient protein-ligand interaction network (AP Layer).

Main Methods:

• The IAP-CFDock framework integrates Iterative Anchor Prediction, Coarse Docking, and Fine-Grained Docking stages.
• An AP Layer network was designed for high-accuracy anchor prediction.
• Performance was evaluated across multiple datasets and compared with existing blind docking methods.

Main Results:

• IAP-CFDock demonstrated superior docking accuracy and generalization capability compared to other methods.
• The AP Layer showed robustness, high accuracy, and flexibility in anchor prediction.
• The framework maintained high inference efficiency across various ligand sizes.

Conclusions:

• IAP-CFDock offers a significant advancement in protein-ligand blind docking, crucial for accelerating drug discovery.
• The proposed method's components are effective, indicating potential for further improvements in computational drug design.