High performance computing support by the Biology IT (information technology) group has saved a College of Liberal Arts and Sciences bioscience research team many hours of work.
Staff in Biology IT (a collaboration of LAS and the College of Agriculture and Life Sciences) created a custom-made server optimized for serial processing. It ran a genome sequence on teosinte (wild maize) by Arun Seetharam, a post-doctoral reserarcher working with Matt Hufford, assistant professor of ecology, evolution and organismal biology. The scientists were studying how the maize genome has evolved during domestication.
A genome assembly connects an organism’s chromosomes through complex computations. However, Hufford said, teosinte genomes “are large and rich in repetitive DNA, which makes them very challenging to assemble and requires a great deal of memory and runtime to complete.” Initial attempts to run the assembly failed until Biology IT staff members John VanDyk, entomology, and Michael Hofmockel and Barry Britt, LAS, became involved.
“We were struggling to find adequate computational infrastructure,” Hufford said. “The high-memory machine constructed by Biology IT has allowed us to complete analyses that previously required over three months in a matter of days.”
Hufford added that the computing power from Biology IT “is allowing us to get beyond our initial assembly of this genome and to start thinking about the comparative, evolutionary analyses that will really tell us something about maize domestication.”
The Biology IT group and LAS Research IT assist faculty to find high performance computing, application development and other research-specific solutions, said Hofmockel, LAS Research IT director.