Edinburgh: Generic array speeds up analysis of biological data 100-fold

Licensing opportunity

Researchers at Edinburgh University have developed a highly parameterised field programmable gate array core for pairwise biological sequence alignment that provides the ability to analyse large biological datasets in a reasonable timeframe.

According to the researchers, the core provides speed-ups in excess of 100 times those of equivalent software implementations running on general purpose computers, and at a fraction of the power consumed. The core could there fore be used for computationally intensive operations that normally require a cluster of computers.

Speeds 150 times as fast have been demonstrated over a Dual-Core AMD OpteronTM processor 2218, using one Xilinx Virtex 4 LX160 -11 FPGA.

A hardware prototype of an FPGA-based web server for biological sequence alignment has been produced.

There are several benefits to hardware implementations of a single FPGA core, when compared to standard desktop software implementations, including computation speed 100 times as fast; 180-fold increase in performance per watt; and a five-fold increase in performance per dollar.

There is support for a variety of sequence types and codes, substitutions matrices, gap penalties and matching tasks. The FPGA core can be used to efficiently implement a range of alignment algorithms and can be can be optimized for specific applications, such as Biological DNA/RNA/Protein sequence alignment in computational biology and in natural language processing, data pattern matching in networking applications and data stream processing in high-end servers.

The university is seeking interest from commercial partners who want to in-license this IP on a non-exclusive basis.

http://www.university-technology.com/details/generic-fpga-core-for-pair…