Project abstract for group dayang

Genomic Selection and Genome-wide Association Studies in Humans and Domestic Animals

Genomic selection using genome-wide SNP markers has rapidly become an accepted technology for genetic improvement in plant and animal speciesand has been recommended as an effective approach to predict disease risks in humans. However, current methodology for genomic prediction and selection focused on additive effects. Based on our methodology for epistasis detection, these researchers have developed methods for genotypic prediction of additive and dominance effects, where dominance effect is a type of non-additive effect. They have successfully developed shared memory parallel computing tool that achieved nearly ideal scalability. They have conducted expensive simulation evaluations of their new methods and computing tools and applied these methods and tools to analyze genomic data in humans, dairy cattle, and swine. For simulation studies alone, they used 70,000 units of computing time. For this coming year, they will have much larger computing tasks: testing the limits of their computing tools, expanding the capability of their computing tool to analyze much larger sample sizes, and conduct data analysis for human and animal genomics data with much larger sample sizes than they have analyzed. In addition, new methods to be developed this year will be much more demanding computationally than current methods, noting that computing time increases approximately cubically as sample size increases. 

A bibliography of publications acknowledging MSI is attached.