Researchers at the U.S. Department of Energy's Argonne National Laboratory have rewritten a widely used genetics analysis package for graphics processors, making a genome-wide study of veterans' data possible. The work centred on SAIGE, a genetics package, and produced a GPU version called SAIGE-GPU, described in a paper in Bioinformatics.
The team found that the original analysis, written for ordinary processors, spent nearly all its time multiplying very large grids of numbers — an operation suited to GPUs. An analysis that would have taken roughly 481,000 CPU-core hours in the original version became tractable after the rewrite, led on the effort by Argonne's Alex Rodriguez.
The Million Veteran Program consortium published results in Science, with Madduri's group among the authors. The study reported on 2,068 traits across 635,969 veterans and flagged 13,672 regions of the genome linked to one or more conditions. About 1,600 of these regions surfaced only because nearly a third of the participants were not of European ancestry, according to the Science paper.
Ravi Madduri, a senior scientist at Argonne, said the work remains at the discovery stage. "This is all preclinical," he said. "This is the discovery phase." He added that things which took 14 days when the Science paper was written now take less than a couple of hours.