High-Throughput parallel blind Virtual Screening using BINDSURF.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Virtual screening (VS) aids clinical research by predicting ligand-target interactions.
• Traditional VS methods often assume a single binding site, overlooking diverse interaction patterns.
• Many VS methods fail to account for ligands binding to multiple, unrelated target sites.

Purpose of the Study:

• To develop a novel VS methodology that addresses the limitations of traditional methods.
• To identify new potential ligand interaction sites on protein surfaces.
• To accelerate the screening of large ligand databases.

Main Methods:

• BINDSURF, a new VS methodology, scans the entire protein surface.
• Utilizes massively parallel GPU hardware for high-speed processing.
• Employs a blind screening approach to identify potential binding hotspots.

Main Results:

• BINDSURF effectively identifies potential ligand interaction sites across the whole protein surface.
• The method is implemented on GPU hardware, enabling rapid processing of large ligand libraries.
• New binding hotspots, not predicted by traditional methods, can be discovered.

Conclusions:

• BINDSURF is an efficient and fast blind VS method for identifying ligand-dependent protein binding sites.
• Leverages GPU architecture for rapid pre-screening of extensive ligand databases.
• Facilitates drug discovery, design, and repurposing by guiding further investigations.