Development of computational tools for the inference of protein interaction specificity rules and functional...

Area of Science:

• Structural bioinformatics
• Computational biology
• Protein interaction analysis

Background:

• Limited protein structures available compared to vast sequence data.
• Underrepresentation of protein complexes in the Protein Data Bank (PDB) despite abundant interaction data from high-throughput experiments.
• Need for computational approaches to bridge the gap between structural and sequence data.

Purpose of the Study:

• To develop computational techniques for extracting and disseminating high-quality data from known protein structures into the protein sequence space.
• To analyze protein complexes in the PDB and predict interaction probabilities for protein sequences.
• To improve the identification and characterization of protein function patterns and surface regions.

Main Methods:

• Analysis of protein complexes in the PDB to calculate interaction probabilities.
• Structural analysis of proteins with PROSITE patterns to build extended patterns, considering conserved residues not obvious in sequence alignments.
• Annotation of protein surface patches using solvent-exposed residues and stringent structural comparison, independent of residue order.

Main Results:

• Development of methods to predict protein-protein interaction probabilities based on structural data.
• Creation of extended sequence patterns by incorporating structural information for proteins with low-selectivity PROSITE patterns.
• Identification of novel sequence patterns through local surface comparison of protein structures.

Conclusions:

• Computational analysis of protein structures and complexes can significantly expand our understanding of protein interactions and functions.
• Integrating structural information improves the accuracy and scope of sequence-based pattern identification.
• The developed methods offer new avenues for discovering protein sequence patterns and predicting interactions.