Chapter 4. Predicting and characterizing protein functions through matching geometric and evolutionary patterns of...

Area of Science:

• Structural biology
• Computational biology
• Bioinformatics

Background:

• Proteins contain internal voids and pockets, some of which are crucial binding sites for molecular interactions.
• Identifying these functional pockets is key to predicting protein functions.
• Similarity in binding pocket characteristics suggests functional similarity.

Purpose of the Study:

• To develop computational methods for distinguishing functional protein pockets from random voids.
• To identify key residues within these functional pockets.
• To predict protein functions at scale by comparing binding surface properties.

Main Methods:

• Geometric computation to identify and analyze protein pockets.
• Methods to identify key residues within functional pockets.
• Bayesian Monte Carlo simulations to differentiate evolutionary pressures of function versus folding.
• Analysis of binding surface residue composition, shape, and orientation.

Main Results:

• Distinguished functional pockets from random voids using computational approaches.
• Identified key residues contributing to pocket function.
• Developed methods for large-scale protein function prediction based on binding surface similarity.
• Utilized Bayesian Monte Carlo to reconstruct evolutionary histories of binding surfaces.

Conclusions:

• Computational analysis of protein pockets and binding surfaces enables accurate function prediction.
• Understanding evolutionary pressures on binding sites aids in identifying functional similarities.
• These methods have implications for drug discovery through analysis of binding pocket characteristics.