A novel method for protein 3D-structure similarity measure based on n-gram modeling

The present paper describes a novel method for measuring structural similarity of proteins in three dimensions. The method gets its roots from computational linguistics and the related techniques for modeling protein structure in string form and pairwise comparison of protein sequences. The method uses n-gram based modeling techniques for capturing regularities in protein structure sequences and joints cross-entropy measures for comparing two protein sequences to do similarity test. In this way, the 3D structure of protein is represented in string form and, then, a similarity test is performed over these sequences. To find an overlap between two protein structures in 3D-space, a superposition task is also applied. In order to confirm the validity of this method, some experiments were performed using a collection of the protein data sets on publicly available servers which showed that the method is efficient.