Friday, November 12, 2010
Technical experts at Carleton’s High Performance Virtual Computing Laboratory (HPCVL) site are helping medical researchers at the University of Ottawa Heart Institute (UOHI) speed up their analytical processes for a research project that could ultimately find drugs to target specific genes associated with certain diseases.
"We’re really excited that we can leverage the computing power of HPCVL for such a significant medical research project,” says UOHI Statistician Robert Davies. “Although it might take five to 10 years to go from the results of our research to clinical trials,Carleton’s computational analysis aids immeasurably to the understanding of common diseases and the genetic susceptibility of these diseases. In essence, their platform is enabling our researchers to solve real-world medical problems a lot faster.”
Davies works on a team headed by Dr. George Wells, and both are involved in Genome Wide Association Studies (GWAS) at the UOHI. Worldwide, these types of studies have identified hundreds of variants associated with various diseases in the last few years. Davies and Wells approached the HPCVL team at Carleton to help them with imputation, one of the most difficult tasks in analyzing their data. Using imputation, a study might go from several hundred thousand to several million genetic variants, drastically improving the power to find genetic variants associated with a specific disease.“We ensured that their sequential software could be parallelized and made to run effectively on our machines,” explains Dr. Peter Taillon, scientific computing specialist at Carleton, “which is an absolutely essential part of accelerating the analytical process.”
This type of computationally demanding analysis is easily handled by the facilities at HPCVL, where it might take only a few days to analyze a GWAS, in contrast to previous efforts that took several months using standard hospital computers. Once genetic variants are discovered, it is up to lab researchers to study how these genes are involved with the disease being studied.
HPCVL is a consortium of universities and colleges, including Carleton, that warehouses state-of-the-art high performance computing hardware and software. Resources help researchers perform complex mathematical calculations that would ordinarily take years to complete on even the most sophisticated desktop computer. HPCVL is a part of the Compute Canada National Platform for HPC.
Carleton’s HPCVL team haS also provided computational support to other leading-edge research projects such as SNOLAB, where results have provided revolutionary insight into the properties of neutrinos and the core of the sun. HPCVL is also lending its computing and storage resources to the ATLAS project, the particle physics experiment at the Large Hadron Collider at CERN in Geneva, Switzerland. This project is studying the basic forces that have shaped our universe since the beginning of time and will ultimately determine its future.