MSPocket: an orientation-independent algorithm for the detection of ligand binding pockets

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Drug Discovery

Background:

• Identifying ligand binding pockets is essential for understanding protein function and designing targeted drugs.
• Existing pocket prediction methods often rely on protein orientation-dependent grid systems, limiting their applicability.
• A need exists for robust, orientation-independent methods for accurate ligand binding site prediction.

Purpose of the Study:

• To introduce MSPocket, a novel program for detecting ligand binding pockets on protein surfaces.
• To develop an orientation-independent algorithm for reliable pocket prediction.
• To evaluate MSPocket's performance against existing state-of-the-art methods.

Main Methods:

• MSPocket utilizes a core algorithm that operates directly on the solvent-excluded surface, avoiding grid-based representations.
• The method's independence from protein orientation was a key design feature.
• Performance was assessed using a standard test dataset and compared against established prediction tools.

Main Results:

• MSPocket achieved 75% accuracy in predicting ligand binding pockets, reaching 92% when considering the top three predictions.
• The program demonstrated comparable performance to leading methods while being protein orientation-independent.
• MSPocket identified distinct pockets compared to other methods, suggesting complementary utility and potential for combined approaches.

Conclusions:

• MSPocket offers a highly accurate and orientation-independent solution for ligand binding pocket prediction.
• The software, available with a graphical user interface and as a PyMOL plugin, facilitates interactive analysis.
• The study advocates for the overlap criterion as a superior metric for evaluating predicted ligand binding pockets.