Parallel PoMSA for Aligning Multiple Biological Sequences on Multicore Computers

The in-depth analysis of the evolutionary of the organisms requires understanding the functional relationship between their biological sequences. Building a homology and evolutionary model on these biological sequences could provide a clear picture of how these organisms developed by the time. Bioinformatics field helps to process, analyze and interpret biological data to automate processing workflow. Multiple Sequence Alignment (MSA) is one of the critical operations which allows a better understanding of the biological relation between sequences by trying to match similar sequences through an automated alignment process. This automation process is provided through a set of well-designed algorithms such as MAFFT, MUSCLE, Clustal-Omega, ...etc. With the explosion of data in different fields, parallel processing through shared memory machines, GPGPU accelerators, and distributed memory systems, becomes a necessity. Therefore, In this work, we propose a multicore version of our sequential MSA algorithm, PoMSA which is called Parallel PoMSA (PPoMSA). The proposed algorithm depends on the accuracy performance gained from the previous proposed sequential version with parallel support that impacts both the accuracy and the execution time performance of the alignment process. The evaluation results show that the proposed algorithm satisfies a higher alignment score compared to existing state-of-art algorithms: Clustal-Omega, MAFFT, and MUSCLE. Moreover, the scalability performance of the proposed algorithm has been evaluated using a different number of cores on a manycore machine and shows strong scalability with larger number of cores.