Real-Coded Genetic Algorithm to Reveal Biological Significant Sites of Remotely Homologous Proteins

摘要

Discovering biological importance from protein structures needs to utilize heuristic approaches. Since three-dimensional (3D) protein structures play a crucial role in biological reactions, comparing new proteins with well-studied proteins is vital for understanding the native functions. A few residues forming in geometrically close positions activate such biological functions. Our goal is to find biological significant sites from protein pairs. Since a pair belonging different protein families has lower global similarity but similar function (so called remotely homologous), we develop a GA-based alignment tool to emphasize small regions of protein pairs geometrically similar.