ROSS : A Rapid Protein Structure Alignment Algorithm

Protein structure comparison is very important in bioinformatics as the function of a protein is determined by its structure. Finding relationships between proteins is very necessary. The databases for proteins are growing rapidly. Therefore there is a strong need for rapid algorithms to handle large number of requests made by biologists. This means that protein structure comparison is an active research area. Relative Orientation of Secondary Structure (ROSS) algorithm for protein structure alignment is proposed. ROSS is a sequential (preserving order) sequence-independent structure alignment algorithm. This work proposes a data reduction and a hierarchical approach for the structure alignment of proteins. The data reduction is done in the pre-processing stage to compute small descriptors for proteins and store them in a repository. This reduction speeds up the process of aligning the structure of a query protein to a large number of proteins. The comparison process is performed hierarchically. The reduced representation depends on the vectors representing the axes of the Secondary Structure Elements (SSE's) and the inter-direction between their centroids. The accuracy of the algorithm is similar to the well known algorithms Secondary-Structure Matching (SSM) and Vector Alignment Search Tool (VAST), used through their websites. The speed of the algorithm is around fifty times that of TM Align algorithm. This speed enables searching for similarities within large datasets instantly. Moreover, the algorithm achieves reduction ratio around 30:1.