Queensland Uni launches biotech data mirror

The University of Queensland has launched a half-terabyte ‘mirror’ of British life sciences data in efforts to facilitate local, data intensive experiments.

The mirror was co-located with the National Computational Infrastructure (NCI) Specialised Facility in bioinformatics.

Together, the facility and mirror comprised 3,144 computing cores in 384 nodes, 11.25 terabytes of RAM and more than 500 terabytes of disk storage.

In advance of the mirror’s launch last Tuesday, data from the European Bioinformatics Institute (EBI) was transferred to the university by tape and made available via EMBL Australia’s web site.

Scientists could download the data to their own research institutions via AARNet’s fibre network, or use it in experiments on the NCI supercomputer.

According to University of Queensland professor Mark Ragan, Australia currently produced only two to three percent of the world’s bioinformatics data.

Most bioscience data sets were stored in facilities in the US, UK and Japan, he said.

Ragan said his own attempts to sequence human cancer genomes were “hugely enriched” by the 98 percent of data housed overseas.

Alpha and beta testers of the EMBL Australia mirror were accessing mirrored EBI data more than 10 times faster than previously, he said.

While the mirror was unlikely ever to serve up all of the EBI’s data – of which there was about 10 petabytes and growing (pdf) – more data would be added in line with Australian research priorities, including crop biotechnology and biodiversity.

Ragan said the mirror used file transfer software Aspera to copy over a terabyte of EBI data a day. “Quite a bit” more data would be sent over on tape within the next two to three months.

The facility had $9.9 million of cash and $3.2 million of in-kind contributions from the Queensland Government, University of Queensland, Bioplatforms Australia, the Australian National Data Services, and the Federal Government.

The University planned to add a “considerably larger amount” of storage to the facility via a tender process later this year, Ragan said, declining to disclose details.