Exhibits, Demos & Posters
Shared Genomics: Accessible High Performance Computing for Genomic Medical Research
Authors
- Mark Delderfield, North West Institute for BioHealth Informatics, University of Manchester, United Kingdom
- Lee Kitching, North West Institute for BioHealth Informatics, University of Manchester, United Kingdom
- Gareth Smith, North West Institute for BioHealth Informatics, University of Manchester, United Kingdom
- David Hoyle, North West Institute for BioHealth Informatics, University of Manchester, United Kingdom
- Iain Buchan, North West Institute for BioHealth Informatics, University of Manchester, United Kingdom
Abstract
Increasingly, genome-wide association studies are being used to identify positions within the human genome that have a link with a disease condition. The number of genomic locations studied means that High Performance Computing (HPC) solutions will have to increasingly be used in the statistical analysis of these datasets. An inevitable reduction in unit cost will bring the genotyping technology within reach of mainstream clinical researchers, who are able to provide the expert interpretation of results but are unlikely to have expertise in HPC and bioinformatics. To ensure that clinical researchers are able to take advantage of genome-wide genotyping technology, we have developed a platform that gives simple access to HPC analysis codes and automatically annotates the results with relevant biological information, whilst hiding as much as possible from the user the HPC infrastructure and workflows processes used. We believe the solution we have developed is amongst the first of its kind. For the HPC analysis codes we developed a parallelized version of an existing statistical genetics application, PLINK, and have successfully deployed it on a 100 processor compute cluster running Windows HPC Server 2008. Access to the platform is via a Windows Forms user interface, though we are currently exploring refactoring the Windows Forms front-end in Silverlight. We have tested our HPC analysis codes and have achieved numerical consistency with the original PLINK codebase for both 32-bit and 64-bit platforms, as well as the expected performance gain when running on multiple cores.
