Applications of High-Performance Computing (HPC) in Plant Breeding, Conservation, and Genetics

Authors: Larson, Steven

Submitted to: Life Sciences Symposium
Publication Type: Abstract Only
Publication Acceptance Date: 3/16/2007
Publication Date: 3/26/2007
Citation: Larson, S.R. 2007. Applications of High-Performance Computing (HPC) in Plant Breeding, Conservation, and Genetics. Life Sciences Symposium.

Interpretive Summary:

Technical Abstract: Increasingly powerful and informative DNA sequencing and genotyping techniques, instruments, and software are being developed and used for a wide variety of applications in agriculture. However, the increasing flux and accumulation of data will also require more HPC facilities and expertise. In cooperation with the USU HPC department, parallel computers were required to test and compare different models of population structure and admixture in key North American range plants using Bayesian clustering of multilocus genotype data. In one study, Bayesian clustering revealed problematic hybridization of native Idaho fescue and introduced sheep fescue grass seed collections that were being promoted for varietal release by the USDA Natural Resource Conservation Service, whereas simple PC-based clustering techniques suggested this germplasm was naturally unique and suitable as a regional variety. Another HPC Bayesian clustering study revealed possible admixture between Lepidium papilliferum endemic to Idaho and nearby forms of L. montanum var montanum in Nevada, Oregon, and Utah. The latter results were considered evidence of relatedness and gene flow between relatively common and rare forms of Lepidium in the recent USFWS decision not to list L. papilliferum as an endangered species, which would have impacted livestock grazing and agriculture in SW Idaho. In cooperation with the University of Illinois W. M. Keck Center for Comparative and Functional Genomics, parallel computers were required to BLAST and align 20,000 basin wildrye and bluebunch wheatgrass expressed gene sequences (ESTs) to ten larger plant EST libraries and the rice genome (430 Mb). The resulting annotations and genome alignments provide a crucial framework for gene discovery in perennial grasses and development of new cultivars adapted to western North America. Thus, HPC had important applications in several recent FRRL projects. More powerful genetic technologies will have important applications in breeding, conservation, genetics, and other keystones of agricultural research and will likely require more HPC facilities and expertise.