Autodock Vina Adopts More Accurate Binding Poses but Autodock4 Forms Better Binding Affinity

Area of Science:

• Computational Chemistry
• Structural Biology
• Drug Discovery

Background:

• Molecular docking is crucial for computer-aided drug design, predicting ligand-enzyme interactions.
• Autodock4 (AD4) and Autodock Vina (Vina) are widely used, open-source docking software.
• Comparing their efficacy is vital for optimizing drug discovery workflows.

Purpose of the Study:

• To evaluate and compare the success rates of Autodock4 and Autodock Vina.
• To assess performance across a diverse set of 800 protein-ligand complexes with experimental binding affinity data.
• To identify optimal docking parameters for each software.

Main Methods:

• Utilized 800 protein-ligand complexes with available PDB structures and experimental binding affinities.
• Performed docking calculations using Autodock4 and Autodock Vina with varied computational settings.
• Analyzed results for correlation between predicted and experimental binding affinities.

Main Results:

• Vina demonstrates faster convergence than AD4.
• AD4 outperformed Vina for 21 targets, while Vina was better for 10 targets.
• Both methods failed to correlate with experimental data for 16 complexes; AD4's 'long' and Vina's 'short' options showed best performance respectively.

Conclusions:

• Neither Autodock4 nor Autodock Vina is universally superior for all protein-ligand complexes.
• Specific target types and chosen software parameters significantly influence docking accuracy.
• Results guide the selection of appropriate docking software and settings for future drug discovery studies.