CSIRO unleashes new bioinformatics centre

CSIRO's gene analysis capability has just been given a massive fillip with the opening of the new CSIRO Bioinformatics Facility (CBF) in Canberra.

The new facility, based around the latest 'blade server' computing platform from Dell, will provide 50 times more grunt than anything previously available, said the joint director of the CBF, Dr Liz Dennis.

"People are already lining up to use the new equipment. Initial projects will include analysing microarray data on wheat, cotton, eucalypt, grape and rice genes. It will give us a much more powerful computing capability, and will be available right across CSIRO," she said.

The hardware itself -- which cost around $375,000 -- is based on a 66-node cluster which offers as much computing power as 130 desktop computers, but in a much smaller space factor.

Gavin Kennedy, the CBF technical project leader, described the CBF opening on March 19th as the "beginning of a whole new venture" for the bioinformatics group.

"The primary goal is to empower CSIRO scientists and to allow them to use bioinformatics readily; and to get them involved in the future development of the CBF," he said.

The installation will be linked to other CSIRO sites around Australia via AARNET, and initially will only be available to CSIRO researchers, although plans are afoot to extend the facility's availability to other establishments such as unis, state agriculture departments, CRCs and similar institutions.

"At the moment we are still working out a basis for they way [the CBF] can work with collaborators. Of course some of the data we will be using is commercially sensitive and we don't want people hacking into it. Security is a big issue," said Dennis. The new facility uses blade server technology which offers a much higher power to space ratio than conventional set-ups, said Kennedy.

"Each chassis carries six blades in a standard rack mount, with a Pentium 3, 1.4GHz chip and 40Gb of storage on each one. We went for the P3 chip instead of the newer P4 as it is integer based, and offered us better sequencing performance than the mathematically-biased P4," he noted.

The hardware will be supported by 1 terabyte of storage, which Kennedy said would probably have to be updated soon. Programs will include BLAST sequencing software, the GENA microarray database, the SEQA sequencing database and other in-house sequencing programs developed by CSIRO.

Two support staff - one system administrator and one software programmer - will keep the virtual wheels turning for the research teams.

"We are expecting this facility to help our teams collaborate 'in the extreme' on their projects," said Kennedy.

"This is a good example of the 'one CSIRO' vision in action. We are all very excited about it," he added.