Detection of multiscale pockets on protein surfaces using mathematical morphology

Area of Science:

• Computational biology
• Structural bioinformatics
• Drug discovery

Background:

• Protein pockets are crucial for small molecule binding.
• Previous methods defined pockets based on probe size, linking probe radius to pocket shallowness.
• Distinct binding molecules exhibit characteristic shallowness distributions.

Purpose of the Study:

• To develop an efficient algorithm for simultaneous detection of deep and shallow pockets (multiscale pockets).
• To introduce a novel program, ghecom, for enhanced binding pocket analysis.
• To assess the performance of ghecom against existing pocket-finding programs.

Main Methods:

• Utilized a 3D grid representation for proteins and probes.
• Applied the theory of mathematical morphology.
• Developed a multiscale approach using various spherical probe sizes.

Main Results:

• The ghecom program demonstrated superior performance in detecting binding pockets compared to four other popular programs.
• ghecom calculates ligand shallowness (R(inaccess)), revealing biases toward specific shallowness ranges for different molecular parts.
• The algorithm efficiently identifies both deep and shallow pockets simultaneously.

Conclusions:

• The ghecom algorithm offers a significant advancement in binding pocket detection and analysis.
• Findings facilitate prediction of molecule-ligand binding likelihood and configuration.
• The ghecom program is accessible via a web server for broader research application.