By Allison Proffitt
June 4, 2013 | Dell announced today the Dell Active Infrastructure for HPC Life Sciences, an infrastructure system optimized for genomic research to begin shipping in July.
The infrastructure includes a scalable 32 node cluster and parallel file system. Workload/cluster management is performed by Bright Cluster Manager (winner of a 2013 Bio-IT World Best of Show award
) The infrastructure comes ready to connect to next gen sequencing systems via Terascala, and storage can connect via InfiniBand or Ethernet. The whole system is “fully integrated at factory for simple deployment.”
“But this is not a sequencer pipeline computational black box appliance,” Tim Carroll, director of global research for Dell, told Bio-IT World this morning. “That’s not the goal of what we’re trying to do… Life scientists, especially, are very focused on their science, as opposed to being focused on the system.”
“Where we give people choice, we did it because we knew those were crossroads where it could be a show stopper for them if they didn’t have a choice. But it’s not five choices there; it’s the two educated choices.”
The product is the result of what Dell calls “customer-inspired innovation.” Dell has been working closely with researchers at TGEN on a genomic application for neuroblastoma. “That’s exactly where the Active Infrastructure for Life Sciences came from: the two-year collaboration with TGEN,” said Carroll. “Even though that was done specifically around these pediatric cancer clinical trials, the applicability of it obviously is so much broader.”
“With diseases like neuroblastoma, hours matter. Our new Dell HPC cluster allows us to do the processing we need to get a meaningful result in a clinically relevant amount of time,” said Jason Corneveaux, a bioinformatician at TGEN in the Dell statement.
TGEN reports a 12-fold improvement in processing power for patient data and genomic analysis in a few hours. The system can process up to 38 genomes per day, Dell predicts. The data to support those numbers is being prepared with TGEN and, “it’s a question of getting that written up,” Carroll says.
“Certainly for Dell, this is the first time that we’ve been able to come to the research community and say, ‘We can give you the platform that will give you increased performance—you’ll get better performance against a pipeline you’re running; less integration time; better supportability; and still do it with some measure of flexibility.’ We’ve been able to hit the four horsemen of what people care about.”
The infrastructure starts at $650,000, what Carroll called the “sweet spot” of what people are spending who have two sequencers. Generally one system will support two to four sequencers. Carroll said Dell is seeing interest from academic groups who are supporting multiple research groups with a few sequencers and small independent groups.
The system has been deployed at TGEN for about five months, Carroll said. "The compute went up very quickly. A lot of the work has been around the storage side of this, that's the really thorny piece. We spent the better part of the spring getting the cluster tuned and optimized."
Now that the infrastructure is available, Carroll believes it will be a key piece in the race to real clinical genomics. The cloud isn't a viable option because of the size of data sets and the regulatory issues, he said. "You're going to have to be able to do this at a local level--regional research center or community hospital. The more we are able to get this technology into a form factor that we can continue to shrink down to get into those environments, the sooner we're going to be able to realize the dream of doing this at a clinical level."