Protein multiple sequence alignment based on secondary structure similarity

Protein multiple sequence alignment is a fundamental task in bioinformatics. Its importance is due to its application in the estimation of the phylogeny tree, prediction of the secondary and tertiary protein structure, function prediction and finding motifs among protein sequences. Obtaining the best solution in multiple sequence alignment (MSA) is an NP-hard problem and whose speed and accuracy researchers have spent three decades trying to improve. In this paper, a novel secondary structure-based approach is presented. At first, sequences are divided into groups according to their secondary structure similarities and then sequences in each group are aligned with each other. At last, alignment results of all the groups are aligned together. Using this method enhances the alignment quality to some extent. A multi-threaded design program was developed to increase the speed. Experimental results are analyzed in terms of qualities and execution times. The results are compared to some commonly used MSA tools.