Flexible Protein Structure Alignment Based on Topology String Alignment of Secondary Structure

Abstract

 Abstract—The paper reports a method for flexible alignment of protein structure. The method in the first phase applies a text modeling technique to obtain an initial superposition of secondary structure elements of two proteins. Then, in the second phase, a step-by-step algorithm is utilized to create flexible alignment between two structures. The method was assessed using a dataset of proteins with macromolecular motions and the results compared with those of the existing flexible alignment methods e.g. FlexProt, FATCAT, and FlexSnap. The results demonstrate that the method have a competitive accuracy in comparison with the other similar methods. chain of AFPs having different number of hinges. HingeProt (11) firstly divides one of the proteins into rigid parts using an approach based on Gaussian-Network-Model, and then, applies MultiProt (12) to align each part with the second protein. RAPIDO (13) uses a flexible aligner that is coupled to a genetic algorithm for the identification of structurally conserved regions. It is capable of aligning protein structures in the presence of large conformational changes. Structurally conserved regions are reliably detected by RAPIDO even if they are discontinuous in sequence but continuous in space and can be used for superpositions revealing subtle differences. Moreover, FlexSnap (14) is a greedy chaining algorithm for flexible sequential and non-sequential alignment of protein structure. The main idea used in the FlexSnap algorithm is to assemble short well-aligned AFPs. FlexSnap has shown a competitive effectiveness in the assessments in comparison with the other state of the art flexible alignment methods by considering non-sequential alignment of the structures. In the recent years, a number of methods have been developed based on linear encoding of protein structure (15)-(18). The methods generally encode protein structure into linear sequences, and then, apply sequence alignment techniques to align two structures. The main idea in development of these methods is to speed up the homology search within a database of protein structures. However, the methods obtain lower accuracy than state of the art geometry based methods. We have developed recently a topology string based method (18) which uses both linear encoding and geometry based schemes to align two protein structures. Thus, the method obtains high running speed as well as linear encoding based methods, while it has a competitive accuracy with geometry based algorithms. Based on the fruitful results of the method, now we extend the scheme for flexible alignment of protein structure.