High Performance FPGA-based Core for BLAST Sequence Alignment with the Two-Hit Method

S. Kasap, Khaled Benkrid, Ying Liu

Research output: Chapter in Book/Report/Conference proceedingConference proceedingpeer-review

6 Citations (Scopus)

Abstract

This paper presents the design and implementation of a high performance FPGA-based core for BLAST sequence alignment with the two-hit method. BLAST with two-hit is a very widely used heuristic biological sequence alignment algorithm, and this paper is the first reported FPGA implementation of it, to our knowledge. The architecture of our core is parameterized in terms of the sequence lengths, match scores, gap penalties, and cut-off and threshold values. It is composed of various blocks each of which performs one step of the algorithm in parallel with the others. This results in a high performance and efficient FPGA implementation, which outperforms equivalent software implementations by one order of magnitude or more. Real hardware implementations show that our core is 52 times faster than equivalent software implementations, on average. Furthermore, the core was captured in an FPGA-platform-independent language, namely the Handel-C language, to which no specific resource inference or placement constraints were applied. Hence, the same code can be easily ported to different FPGA families and architectures.
Original languageEnglish
Title of host publication8th IEEE International Conference on BioInformatics and BioEngineering
PublisherIEEE
Pages1-7
Number of pages7
ISBN (Print)978-1-4244-2844-1
DOIs
Publication statusPublished - 8 Dec 2008
Externally publishedYes
Event8th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2008 - Athens, Greece
Duration: 8 Oct 200810 Oct 2008

Conference

Conference8th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2008
Country/TerritoryGreece
CityAthens
Period8/10/0810/10/08

Fingerprint

Dive into the research topics of 'High Performance FPGA-based Core for BLAST Sequence Alignment with the Two-Hit Method'. Together they form a unique fingerprint.

Cite this