Low-cost hierarchical memory-based pipelined architecture for DNA sequence matching

A novel design on faster processing of DNA sequences base on pipelined structure has been presented in this paper. The focus of this paper is to design and develop the hardware architecture to meet the requirement of such processing. In the proposed architecture, instead of comparing the whole DNA search string at once, the entire comparison process is split into four phases using four-stage pipeline structure, to boost up the speed of processing by exploiting the inherent temporal parallelism in the search process. Moreover, we reduced memory requirement by 96.86% by arranging DNA sequences in an intelligent manner compared to the straight and conventional method of storing DNA sequences using Content Addressable Memory (CAM). This effectively enhances system throughput by searching more number of DNA sequences against a single DNA sequence in traditional design. It further reduces system size, cost and power by replacing CAM using Random Access Memory (RAM). It is observed that, only a fraction (0.39%) of the total memory available in Stage 2 is activated, while 0.0015% of the total memory is activated in Stage 3 and a very low fraction (0.000006%) of the total memory is activated in Stage 4, during the comparison of a DNA sequence in the proposed four-stage pipelined structure. This effectively lowers the system cost, power consumption, the heat dissipation, and space requirement.