Bioinformatics research field is becoming an increasingly important field of research. The reason for that is the increasing interest in studying the structure and the function of DNA, and correlating this information with diseases.
DNA and Protein database sequence alignment such as Smith-Waterman Algorithm is one of the most important applications in Bioinformatics. This algorithm guarantees to return the optimal alignment of two sequences. The first one is called Query and the second one is called Database. These applications need to process large amounts of data which may reach more than 11 GB size and that may take hours of mainframe time to get the optimum solution.
Different methods are used to improve the execution performance and to speed up the whole process using a GPGPU (General Purpose Graphical Processing Unit), a CPU, or configurable configurable gates such as FPGA. We introduce a novel and efficient technique to improve the performance of the Smith-Waterman algorithm using a unique platform. Our implementation improves the average performance by 127% (i.e. 2.2x), and shows peak performance improvement of 180% (i.e. 2.8x) in comparison to the state-of-the-art implementation
PIs: Dr. Talal Bonny and M Affan Zidan