How and Why Does Knowledge-Based Biased Docking Improve Molecular Docking Performance?

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Biophysics

Background:

• Protein-ligand interactions are crucial in biology and drug discovery.
• Traditional docking methods often struggle to accurately predict binding poses and affinities.
• The influence of solvent structure on protein-ligand interactions has been historically underestimated.

Purpose of the Study:

• To review the development and application of the biased docking strategy.
• To demonstrate how solvent structure can improve molecular docking predictions.
• To provide guidance on implementing biased docking in current software.

Main Methods:

• Molecular simulations using mixed solvents (e.g., ethanol, phenol, isopropanol).
• Development and application of the biased docking strategy.
• Analysis of protein-solvent and protein-ligand interactions.
• Implementation guidance for existing docking software.

Main Results:

• Protein-solvent interactions in mixed solvents mimic experimental protein-ligand complexes.
• Biased docking significantly improves pose prediction and scoring in docking.
• The strategy is applicable to protein-ligand docking, virtual screening, protein-protein docking, and metalloprotein docking.

Conclusions:

• Biased docking is a powerful strategy for enhancing molecular docking accuracy.
• Understanding protein-solvent interactions is key to improving computational predictions.
• This approach offers valuable insights for developing future AI and Machine Learning-based methodologies.