GPU-Accelerated Drug Discovery with Docking on the Summit Supercomputer: Porting, Optimization, and Application to...

Area of Science:

• Computational chemistry and molecular modeling.
• Drug discovery and development.
• High-performance computing applications.

Background:

• Protein-ligand docking is a crucial in silico method for screening drug candidates.
• Experimental drug screening is costly and time-intensive, necessitating high-throughput computational approaches.
• Existing docking tools often lack optimization for high-performance computing (HPC) resources.

Approach:

• Porting and optimizing the AutoDock-GPU program for the Summit supercomputer.
• Leveraging HPC resources to enhance the speed and scale of docking calculations.
• Validating the optimized program's performance and accuracy.

Key Points:

• AutoDock-GPU was successfully adapted and optimized for the Summit supercomputer.
• The optimized program enables large-scale, high-throughput virtual screening of drug compounds.
• This approach significantly narrows the search space for experimental drug discovery.
• The tool was applied to screen compounds targeting SARS-CoV-2 proteins.

Conclusions:

• Optimized AutoDock-GPU on HPC platforms can accelerate drug discovery campaigns.
• This computational strategy is vital for efficiently identifying potential therapeutics for pandemics like COVID-19.
• HPC-enabled docking provides a cost-effective and rapid alternative to traditional screening methods.