Constructing a Foundational Platform Driven by Japan's K Supercomputer for Next-Generation Drug Design.

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Rising pharmaceutical research and development (R&D) costs pose a significant challenge globally, particularly in Japan due to its rapidly aging population.
• The escalating R&D expenses impact not only the pharmaceutical industry but also the broader healthcare system.

Purpose of the Study:

• To investigate the utility of the K supercomputer for large-scale drug discovery initiatives.
• To evaluate custom-tailored computational frameworks for big data and simulation-based drug discovery.

Main Methods:

• Implementation of primary and secondary computational methods optimized for the K supercomputer's 88,128 compute nodes/CPUs.
• Execution of virtual screening for approximately 19 billion compound-protein interactions.
• Application of a computationally intensive binding free energy algorithm.

Main Results:

• Accurate prediction of compound-protein interactions validated against experimental data.
• Computationally derived binding free energies showed considerable accuracy when compared to experimental data.
• Demonstrated the feasibility of large-scale computations for drug discovery on the K supercomputer.

Conclusions:

• The K supercomputer, with its tailored frameworks, offers a powerful platform for accelerating drug discovery through big data and simulations.
• The implemented computational approaches are effective for large-scale drug discovery and applicable to similar high-performance computing infrastructures worldwide.
• This study provides a scalable framework for addressing the challenges of rising pharmaceutical R&D costs.