ApoDock: Ligand-Conditioned Sidechain Packing for Flexible Molecular Docking

Abstract

Molecular docking is a crucial technique for elucidating protein-ligand interactions. Machine learning-based docking methods offer promising advantages over traditional approaches, with significant potential for further development. However, many current machine learning-based methods face challenges in ensuring the physical plausibility of generated docking poses. Additionally, accommodating protein flexibility remains difficult for existing methods, limiting their effectiveness in real-world scenarios. Diffusion based models has already show a good solution of those problems, such as Alphafold3(AF3). Herein, we present ApoDock, a modular docking paradigm that combines machine learning-driven conditional sidechain packing based on protein backbone and ligand information with traditional sampling methods to ensure physically realistic poses. With accurate sidechain packing and physical based pose sampling, ApoDock demonstrates competitive performance across diverse applications, highlighting its potential as a valuable tool for protein-ligand binding studies and related applications.