Evaluations of molecular docking programs for virtual screening

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Structure-based virtual screening (SBVS) is crucial for identifying drug candidates.
• Numerous molecular docking programs exist, making tool selection challenging.
• Understanding program performance is essential for effective SBVS.

Purpose of the Study:

• To evaluate and compare the screening performance of three prominent molecular docking programs: DOCK, AutoDock, and GOLD.
• To provide data-driven recommendations for enhancing molecular docking program efficacy.
• To validate screening performance using established and novel metrics.

Main Methods:

• Evaluated DOCK, AutoDock, and GOLD against 116 diverse protein targets.
• Employed a rigorous validation approach using two novel standards and traditional enrichment rates.
• Conducted 1000 docking runs per program with three ligand conformations for statistical robustness.

Main Results:

• Each program exhibited unique strengths and weaknesses in screening performance.
• DOCK consistently outperformed AutoDock and GOLD in enrichment rate measurements.
• Performance variations were observed across different target proteins and docking programs.

Conclusions:

• DOCK shows significant promise for structure-based virtual screening applications.
• Recommendations are provided to improve the screening capabilities of molecular docking software.
• This comparative analysis aids researchers in selecting appropriate tools for drug discovery pipelines.