Based on the open source Mitrion-C Open Bio Project, Mitrionics is providing a turnkey Mitrion-accelerated version of NCBI BLAST. The package is available through a subscription to updates and includes installation support. Mitrion-Accelerated BLAST accelerates the BLAST-n nucleotide search program of the NCBI BLAST application up to 60x compared to a 2.8GHz AMD Opteron Processor*

 

SGI RASCThe Mitrion-Accelerated version of NCBI BLAST is available now and runs on SGI® RASC™ hardware. It comes pre-installed on the SGI® RASC™ Appliance for Bioinformatics. This is the world’s first turnkey bioinformatics appliance based on FPGA technology and open source software. The solution is designed to break through query logjams that result from next-generation sequencing instruments and multiple genome studies using NCBI BLAST-n. And since the Mitrion-Accelerated BLAST source code is released as open source, users are at liberty to tweak the application to fit their needs.

 

The SGI® RASC™ Appliance for Bioinformatics – Featuring Mitrion-Accelerated BLAST-n, is far more cost- and energy- efficient than traditional rack server deployments, because it relies on FPGA technology for the bulk of BLAST-n query processing. The Appliance consumes up to 90 percent less power than an Opteron- or Xeon-based rack system. This equates to lower cost of ownership and longer asset life.

 

The Mitrion-C Open Bio Project
The purpose of the Mitrion-C Open Bio Project is to accelerate key bioinformatics applications by porting their performance critical parts to Mitrion-C and running these on the Mitrion Virtual Processor in FPGAs. In addition to the BLASTN program, work has been started on accelerating BLASTP. Plans are then to continue with the Translated BLAST programs, Smith-Waterman and Hidden Markov Models (based on Sean Eddy’s HMMer package), depending on the priorities of the development community.

 

*) Searching the Unigene Human and Refseq Human databases (6,733,760 sequences; 4,089,004,795 total letters) using a small subset of the Human Genome U133 Plus 2.0 Array probe set from Affymetrix (3534 sequences; 88,350 total letters):

  • SGI Altix/RC100 using 1 FPGA: 0h 09m 29s
  • 2.8GHz Dual Core AMD Opteron Processor 8220 SE using 1 core: 9h 30m 7s

Speedup: 60.1x