In the academic world, grid computing stirs the imagination of life sciences researchers with visions of accessing biological databases and computing resources across the globe. So far, however, concerns about data security and investing in an immature technology have discouraged scientists in industry from participating in grid projects as enthusiastically as their academic counterparts.
But at least in the UK, that may be changing. In one effort, centered in Belfast, Northern Ireland, researchers at Queens University Belfast have enlisted two local biotech firms — Fusion Antibodies and Amtec Medical — to participate in developing GeneGrid, a year-old project to create a “virtual bioinformatics laboratory.” Although the companies don’t have to contribute funding — the collaborative project was awarded an initial £500,000 grant from the UK e-Science program — scientists from the two companies say they’re helping design GeneGrid’s interface and capabilities as well as serving as guinea pigs by testing the initial versions of the system.
GeneGrid itself is a product of the Belfast e-Science Centre, one of eight regional grid computing centers of excellence across the UK. Led by Ron Perrott, a professor of software engineering at Queens, the Belfast effort is attempting to differentiate itself from other grid computing projects by focusing on industrial use cases, says Paul Donachy, the commercial director for the Belfast e-Science Centre.
Donachy says GeneGrid is taking an industrial standpoint by soliciting input from Amtec and Fusion on how they’ll want to use the grid. “Our remit was to identify basic ‘problems’ for the GeneGrid to solve, and to identify public-access genetic databases for plugging into the GeneGrid,” writes Amtec research scientist Shane McKee in an e-mail.
More specifically, GeneGrid consists of middleware designed by the Belfast group that relies on the Open Grid Services Architecture model. Donachy says the middleware can pull in applications along with various public and private data sets to complete a user-specified bioinformatics computation. Through the grid, Donachy says users will be able to access computing resources at the other seven e-Science centers in the UK, as well as at the San Diego Supercomputing Center. All traffic is encrypted to alleviate concerns about data security, Donachy adds.
As this magazine went to press, scientists at Amtec and Fusion were about to test-drive a second version of GeneGrid in the hopes of working out the bugs before the end of the year. The new version, says Paul Kerr, Fusion’s director of business development, is expected to be more user-friendly, as the initial version of GeneGrid “required a lot of code up front.” The user interface, he adds, “was the last thing to be sorted out.” By the spring, Kerr plans to install GeneGrid as an integral part of the company’s bioinformatics infrastructure.
Ultimately, Fusion hopes to use GeneGrid to help perform massively parallel Blast searches to identify novel drug targets, Kerr says, followed by calculations to predict protein structures and solubilities for promising drug targets. Fusion, also based in Belfast, has plans to seek out its own proprietary drug targets — in addition to performing protein expression and antibody production services for other pharma and biotech companies.
Amtec, a developer of medical devices with a tradition of bioinformatics expertise, is taking a more wait-and-see approach. Writes McKee: “Whether [GeneGrid] ends up making a difference to medical devices is open to question, but there will be market opportunities for genetic data analysis, and Amtec’s extensive experience in this area will prove very useful.”