Enriching the gold dust:
extreme-value based genome-wide association in the post-GWAS era
Dalin Li
Division of Biostatistics
University of Southern California
Friday, September 11, 2009, 12:30–1:30 pm
GEMS classroom, 3rd Floor in
Shriner's Building
Coffee, tea, and cookies will be provided
Abstract
Variants identified in recent genome-wide association studies (GWAS) based on the common-disease common-variant hypothesis are far from fully explaining the heredity of complex traits. To detect rare variants, very large sample size is required and the cost would be forbidding, particularly when sequencing might be necessary. Here we propose the extreme-value based GWAS in which only individuals with extreme phenotypes are recruited for whole-genome genotyping or sequencing. Our investigation shows that this design is highly efficient for the rare variants, with the sample size required to achieve the same power can tens or even hundreds of times less than traditional designs. We propose an analysis framework for extreme-value design based on the maximum likelihood theory. A corresponding power calculation approach and a guideline for optimizing this design conditional on phenotyping/genotyping cost ratio are further proposed.
This is the joint work with David V. Conti.