Text modeling approach for protein 3D structure similarity measurement

Abstract

Protein structure similarity measure is a very important tool to highlight the similarities and differences between protein structures. It has wide applications in protein structure analysis and classification, structure-based drug design, phylogenetic analysis and protein structure predictions which have attracted tremendous attention and have been broadly studied within the past decade. It is known that the protein structure highly indicates its functionality and the potential interactions with the other protein structures. For two given proteins, if the sequences are similar then their evolutionary relationship is obvious. Otherwise, the 3D-structure of proteins, due to structural and functional behavior placed on them, are much more evident than protein sequences. As a result, the structural similarity of proteins should be used to distinguish the differences among various proteins functionalities and their evolutionary relationships. Whenever a new protein structure is discovered, it is desired to find the structural similar proteins to predict its functions and properties.