Comprehensive assessment of flexible-ligand docking algorithms: current effectiveness and challenges.

Area of Science:

• Computational chemistry
• Drug discovery and development
• Molecular modeling

Background:

• Protein-ligand docking is crucial for drug discovery.
• Flexible-ligand docking algorithms have advanced, but rigid-ligand docking remains computationally efficient.
• Assessing the comparative effectiveness of flexible versus rigid ligand docking is essential.

Purpose of the Study:

• To compare flexible-ligand docking with rigid-ligand docking for virtual screening and pose prediction.
• To evaluate three representative docking algorithms: global optimization, incremental construction, and multi-conformer docking.
• To identify challenges in current docking algorithms, particularly in scoring functions.

Main Methods:

• A comprehensive assessment of flexible-ligand versus rigid-ligand docking was performed.
• Three docking algorithms were tested on the Directory of Useful Decoys dataset.
• The study analyzed performance in both virtual screening enrichment and pose prediction accuracy.

Main Results:

• Flexible-ligand docking did not yield statistically significant improvements in virtual screening enrichment compared to rigid-ligand docking.
• All tested docking programs significantly improved pose prediction success rates when ligand flexibility was considered.
• The effectiveness of flexible-ligand docking was better in pose prediction than in virtual screening.

Conclusions:

• Current docking algorithms face greater challenges in ranking than in docking itself.
• Considering ligand flexibility is justified in virtual screening, especially for more flexible ligands.
• Addressing scoring challenges, including ligand conformational entropy, is critical for improving docking accuracy.