Protein Structure Modeling in Two-Level Topology Strings for Structure Comparison

Abstract

Biological data representation in textual sequences has opened a new hybrid research area in protein structure analysis. The paper presents a novel method for structural comparison and alignment of proteins based on text modeling techniques. The method firstly models the geometry of protein secondary and tertiary structure in two-level topology strings and then, employs n-gram modeling and cross-entropy concept from computational linguistics approach to compare strings and align structures. The results of the experiments confirm the validity and efficiency of the proposed method. Moreover, conceptual simplicity of the method indicates applicability and preference of the method in comparison to the other related techniques.