Molecular dynamics-based virtual screening: accelerating the drug discovery process by high-performance computing.

Area of Science:

• Computational chemistry and drug discovery
• Biopharmaceutical innovation leveraging advanced computing

Background:

• High-performance computing (HPC) is a critical national strategic technology.
• A key hypothesis suggests HPC can accelerate biopharmaceutical innovation.

Purpose of the Study:

• To experimentally demonstrate HPC's capability to accelerate biopharmaceutical innovation.
• To evaluate the effectiveness of molecular dynamics-based virtual screening (MDVS) with HPC.

Main Methods:

• Utilizing molecular dynamics-based virtual screening (MDVS) on HPC systems.
• Comparing computational times of MDVS on HPC versus traditional PC servers for large compound libraries and simulations.

Main Results:

• HPC significantly accelerates MDVS, reducing years of computer time to feasible durations.
• State-of-the-art HPC systems demonstrated up to 600x speedup compared to an eight-core PC server for drug target screening.
• Optimized GPU/CPU architecture design can effectively reduce HPC costs.

Conclusions:

• HPC is a powerful tool for accelerating biopharmaceutical innovation through MDVS.
• The primary limitation for further advancements in virtual screening is the communication cost in parallel computing.