Improving Blind Docking in DOCK6 through an Automated Preliminary Fragment Probing Strategy

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurately predicting small molecule binding sites on protein surfaces is crucial for drug discovery.
• Current methods like blind docking and cavity detection-guided docking have limitations in terms of speed and accuracy.
• A hybrid approach is needed to balance efficiency and precision in virtual screening.

Purpose of the Study:

• To develop and validate BLinDPyPr, an automated pipeline for hybrid blind docking.
• To improve the efficiency and accuracy of predicting small molecule binding sites and conformations.
• To bridge the gap between conventional blind docking and cavity detection-guided docking methods.

Main Methods:

• Developed BLinDPyPr, an automated pipeline integrating FTMap and DOCK6.
• Utilized FTMap to generate probe clusters, converted them into DOCK6 spheres for binding region identification.
• Implemented a hybrid blind docking strategy for simultaneous probing and selection of binding pockets based on ligand properties.

Main Results:

• Achieved pose prediction success rates of 45.2-54.3%, comparable to site-specific docking.
• The BLinDPyPr pipeline demonstrated a 50% reduction in time compared to conventional blind docking with DOCK6.
• Successfully identified and prioritized potential binding regions based on ligand-specific properties.

Conclusions:

• BLinDPyPr offers a practical and efficient solution for virtual screening of large molecule libraries.
• The hybrid docking strategy enhances accuracy while significantly reducing computational time.
• This method provides a valuable tool for accelerating drug discovery and development processes.