Gene Enrichment in Genome-Wide Association Studies

• Activity Number: 515
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Stephen Aaron Stanhope*+ and Mark Abney
• Companies: The University of Chicago and The University of Chicago
• Keywords: Genome-wide association study ; Gene enrichment

Abstract

In a genome-wide association study (GWAS), hundreds of thousands of markers are genotyped across hundreds of cases and controls, and each marker is tested for disease associations. After controlling for multiple comparisons, it can be difficult to detect loci with small effects on disease risk. It is therefore attractive to limit the number of tests by analyzing groups of related markers. We describe an enrichment study of hypertension in the Hutterites, a population related through a 13-generation pedigree. We aggregate score statistics from a GWAS into gene scores using the array annotation and a scoring function. Because of linkage disequilibrium, under the null of no markers within a gene being associated with the disease, the distribution of the gene score is that of a function of correlated chi-squared distributed random variables. We obtain p-values by sampling normal random vectors with dependencies derived from observed patterns of variation in gene scores. This strategy yields correct type I errors and increases power. We identify a number of genes associated with hypertension in the Hutterites, including several previously identified in other populations.