Local Structure Comparison of Proteins

Abstract

Protein local structure comparison aims to recognize structural similarities between parts of proteins. It is an active topic in bioinformatics research, integrating computer science concepts in computational geometry and graph theory with empirical observations and physical principles from biochemistry. It has important biological applications, including protein function prediction. In this chapter, we provide an introduction to the protein local structure comparison problem including challenges and applications. Current approaches to the problem are reviewed. Particular consideration is given to the discovery of local structure common to a group of related proteins. We present a new algorithm for this problem that uses a graph-based representation of protein structure and finds recurring subgraphs among a group of protein graphs.