Optimization methods for virtual screening on novel computational architectures

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Virtual screening (VS) methods are crucial in drug discovery, differing in receptor/ligand modeling and screening approaches.
• Current VS methods screen large chemical databases (e.g., ZINC), but computational time increases with database size and model accuracy.
• Limitations in computational resources hinder the application of detailed, high-accuracy models in VS.

Purpose of the Study:

• To review recent trends in virtual screening modeling techniques.
• To explore the impact of novel computational architectures on VS performance.
• To discuss the current state and future trends of advanced computing in drug discovery.

Main Methods:

• Review of existing virtual screening methodologies.
• Analysis of computational time requirements for different VS approaches (fast docking vs. molecular dynamics).
• Investigation of novel hardware platforms and advanced computing trends.

Main Results:

• Novel computational architectures offer order-of-magnitude improvements in VS processing speeds.
• Increased computer power at lower cost enhances VS method performance.
• Better quality and quantity of conclusions from VS screening are achievable.

Conclusions:

• Advancements in computational power and novel architectures are overcoming VS bottlenecks.
• The integration of advanced computing with improved modeling techniques is revolutionizing drug discovery.
• Future VS applications will benefit from enhanced speed, accuracy, and scope.