Authors: Kaushik Roy; Asish Mukhopadhyay; Gilbert Cole
Addresses: Bell Mobility, 100 Wynford Drive, North York, Ontario M3C 4B4, Canada ' School of Computer Science, University of Windsor, Windsor, ON, N9B 3P4, Canada ' School of Computer Science, University of Windsor, Windsor, ON, N9B 3P4, Canada
Abstract: Multiple structure alignment (MStA) is a fundamental tool for correlating the structural similarity of proteins with their functional similarity. A number of algorithms are extant, MUSTANG, POSA, MultiProt, CE-MC to name a few. In this paper we propose a new algorithm, MASCOT. This uses the DSSP program to map a protein structure into a DSSP-sequence, reducing the structural alignment problem to a sequence alignment problem. We have used a center-star approach to select a center-protein with respect to which to create an alignment. The root mean square deviation (RMSD) has been used to measure alignment quality for a large variety of datasets. We compared the execution times of our algorithm with the well-known algorithm MUSTANG for all the tested alignments. MASCOT outperformed MUSTANG on all but one sample. Another measure (Alignment Accuracy), was used to compare the performance of MASCOT and MUSTANG for protein structures drawn from the manually curated database HOMSTRAD.
Keywords: structural bioinformatics; protein structure alignment; computational biology; algorithms.
International Journal of Bioinformatics Research and Applications, 2021 Vol.17 No.2, pp.135 - 157
Accepted: 30 Oct 2018
Published online: 18 Apr 2021 *