Fast, accurate, and reliable molecular docking with QuickVina 2.

Area of Science:

• Computational chemistry
• Drug discovery and development
• Bioinformatics

Background:

• The increasing size of drug-fragment databases necessitates faster molecular docking tools for high-throughput screening.
• AutoDock Vina (Vina) is a popular, parallelized docking software.
• QuickVina 1 (QVina 1) accelerated docking but compromised accuracy at low exhaustiveness.

Purpose of the Study:

• To introduce QuickVina 2 (QVina 2), an improved molecular docking tool.
• To combine the speed of QVina 1 with the reliability of Vina.
• To evaluate QVina 2's performance in terms of speed and accuracy.

Main Methods:

• QVina 2 was tested on the PDBbind 2014 core set.
• Performance was compared against Vina and GOLD using standard exhaustiveness levels.
• Key metrics included speed acceleration, correlation coefficients, and Root Mean Square Deviation (RMSD).

Main Results:

• QVina 2 achieved an average of 2.30-fold acceleration over Vina at Vina's default exhaustiveness (8).
• High correlation coefficients (0.967 for the first mode, 0.911 for all modes) were observed.
• QVina 2 showed improved accuracy over GOLD, with 63.1% of cases having RMSD < 2 Å compared to GOLD's 56.9%.

Conclusions:

• QVina 2 offers a significant speed enhancement over Vina while maintaining comparable reliability.
• The tool demonstrates competitive accuracy, outperforming GOLD in RMSD metrics on the tested dataset.
• QVina 2 represents a valuable advancement for efficient molecular docking in drug discovery.