Parallelization of protein clustering algorithm using OpenMP

Abstract

Proteins are the building blocks of all living organisms and its analysis can help us to understand the bimolecular mechanics of living organisms. Protein clustering attempts to group similar protein sequences and has diverse applications in bioinformatics. However, this operation faces various computational challenges because of dependency on complex data structures, high memory usage and irregular memory access patterns. In genome studies, the time consideration for alignment is also an important parameter and should be minimized. Conventional solutions have rather been unsuccessful in achieving decent runtime performance because these algorithms are designed for serial computation which means that they use a single processor to perform computations. These algorithms can be improved upon by modifying them to use multiple processing elements. The purpose of this research is to modify existing protein clustering algorithm and apply parallelization techniques on them in order to optimize protein sequencing operation for faster results without sacrificing accuracy.