Department of Statistics Seminar
North Carolina State University

presents

Dr. Timothy Thornton

University of Chicago

"More Powerful Quasi-Likelihood Score Tests for Genetic Association in Case-Control Studies"

ABSTRACT

Case-control studies have been extremely valuable in evaluating associations between candidate genes and complex diseases. Traditional case-control studies use unrelated subjects and compare allele or genotype frequencies of the cases and the controls at genetic markers. When affected related individuals are used in association studies, the power to detect an association is increased since affecteds with affected relatives have a higher expected frequency of the alleles that increase susceptibility for a genetic trait than do affected individuals that do not have affected relatives. When related individuals are used in a study, the correlations among the relatives must be taken into account to ensure validity of the test, and consideration of these correlations can also improve power. To test for allelic association with a binary trait, we propose a new test, the M_QLS test, that is an extension of the quasi-likelihood score test of Bourgain et al. (2003) and which takes advantage of the fact that affected individuals that have affected relatives are more likely to have the predisposing variants than individuals that do not have affected relatives. One of the motivations for using the M_QLS test is that for any arbitrary set of outbred individuals, the M_QLS statistic has maximal non-centrality parameter in a general class of linear statistics, for all 2 allele disease models, as the effect size tends to zero. Allelic association tests essentially treat the alleles, instead of the individual, as the sampling unit and are valid under the assumption of Hardy-Weinberg equilibrium (HWE). We introduce a new genotypic test for association that is an extension of the genotypic test of X. Wu (unpublished Master's thesis) and which is valid in the absence of HWE. Simulations are performed to compare the power and type I error of the new tests with those of competitors. We apply the allelic methods to analyze data on asthma-related phenotypes in a complex Hutterite pedigree and an alcoholism related phenotype in a sample of moderate-size outbred Caucasian pedigrees from GAW 14 COGA data. We apply the genotypic methods to test for association between hypertension and markers at a candidate gene in a sample of moderate-size African-American pedigrees.

Tuesday, April, 18, 2006

9:00 - 10:00 am

208 Patterson Hall

Refreshments will be served on 206 Patterson Hall at 8:30 am.