Green Data Center Rackspace Image
»
UA Research Computing an Asset to Epilepsy Studies
UA Research Computing an Asset to Epilepsy Studies
Researchers Krishna Veeramah, Laurel Johnstone, and Mike Hammer.

Early onset epilepsy, a neurologic condition characterized by recurrent seizures beginning in early childhood, can induce developmental disabilities and in some cases result in intellectual disability or even death in affected children. Recently, UA research scientist Mike Hammer and his colleagues at Arizona Research Labs, research specialist Laurel Johnstone and post-doctoral researcher Krishna Veeramah, explained how technology plays a key role in helping them progress toward identifying the genetic source of the condition.

Genomes may vary from each other by millions of mutations, which is beneficial from the perspective of genetic variation. Such mutations are unavoidable but are rarely lethal; however, a mutation that alters the function of a resultant protein can bear potential complications in development and it’s those mutations that Hammer’s group is analyzing. In previous studies, it was found that childhood epilepsy may stem from de novo mutations, those unique mutations unique to found only in the child’s sequence. Based on those findings, Hammer and his colleagues work with families of epileptic children who volunteer to be studied.

 “Right now, we are sequencing the exome, which are small portions accounting for about 1 to 2 percent of the entire genome,” says Hammer, “and if at some point we do have to sequence the entire genome, we can.” Their decision to use whole exome sequencing focuses their data to portions of the genome that may be related to epilepsy.

These steps are accomplished using UA’s High Performance Computing (HPC) resources that enable several processing jobs to be accomplished simultaneously. “For this project we run things on a single node on HTC using 12 processors and 24 gigabytes of memory. We could do that on our local server, but it would limit us to doing one job at a time,” Johnstone says of the resources she uses in her calculations. “With HPC, I can run multiple things in parallel without disrupting the entire lab that is also working on our server.”

Hammer’s group use the standard, no cost, allocation of 10,000 cpu-hours per month per user to run their research calculations. Genomics data and analysis results requires large amounts of storage, so the Hammer group “rents” 14TB of storage on the HPC systems to store their data and results.

DNA sets gathered from participating families are then processed using Burroughs-Wheeler Alignment, a memory-efficient program that aligns the sequences overnight before analysis can begin. To run the subsequent pipeline of processes requires what Krishna regards as a synergy of theory and technology. The bulk of the computational work is managed by Johnstone who collaborates with Veeramah on devising the strategies to process their collected genetic data from sequencing to later alignment and analyses of significant de novo variant sites in an epileptic child’s sequence.

So far, their lab has compared about 50 people, analyzing one family at a time. Just how useful the university’s HPC systems are to their work, Johnstone says, is apparent in processing times. Tasks that in the past would have taken up to 11 days are now reduced to a mere 24 hours. With such resources for increased efficiency, the team is optimistic about their current research and is hopeful their contributions will make in the overall efforts to someday cure epilepsy.

In addition to their continuing work on epilepsy, Hammer, Johnstone, and Veeramah are also using HPC resources for a genome mapping project on gibbon apes. The goal is to simulate the genome to make inferences regarding the genetic background of great apes by running multiple sequences of ape DNA and looking for patterns throughout. While the gibbon project, as well as their work on epilepsy, is evidence of the progressive scientific efforts being made by university researchers, much still lies ahead. On the future of technology in research, Hammer says, “We think that this is the way medicine is going, so we want to position ourselves for success and having adequate computing power is obviously an asset in preparing for that.”