An FPGA-based parameterised and scalable optimal solutions for pairwise biological sequence analysis

Abstract

This paper presents the design and implementation of the computationally intensive Smith Waterman and Needleman-Wunsch algorithms on the Virtex-5 (XC5VLX110) FPGA (Field Programmable Gate Array). A parameterisable architecture of the processing element (PE) for the pairwise biological sequence analysis is presented. We evaluate our core's performance in terms of speed up performance and normalised cell update per second (CUPS) against various computing platforms including our previous Graphics Processor Unit (GPU) and other FPGA implementations. Our efficient hardware design produces the highest normalised throughput ever reported in literature.