DNA sequence alignment: A review of hardware accelerators and a new core architecture

摘要

Deoxyribonucleic Acid (DNA) sequence alignment is essentially a way of comparing two or more DNA sequences with aim to find regions of similarities among them. The Smith-Waterman (SW) algorithm is a local alignment algorithm which is able to identify mutation in DNA sequences. However, the aforementioned algorithm tends to be slower in computation of long DNA sequences. Over decades ago, Field Programmable Gate Arrays (FPGAs) play an important role in DNA sequence alignment. Moreover, pipelining technique is also a well-known method used to speed-up the performance of hardware design. Systolic array (SA)-based DNA sequence alignment architecture reduces execution time of alignment matrix computation from quadratic to linear time complexity. In this paper, existing FPGA-based sequence alignment core architectures will be discussed followed by proposal for a new SA-based DNA sequence alignment core architecture. The design was synthesized on the Xilinx Spartan-3E XC3S1600E-FG3205. Results showed that the developed core architecture is 1.2× faster in speed as compared to other reported FPGA-based designs.