This application is a continuation of our work on the development and implementation of computer-based methods in genetic epidemiology. Genome scans of common disease are now published regularly, and many yield strong evidence that multiple loci influence disease expression. The subject of how locus interaction affects linkage analysis of diseases caused by multilocus systems has gone from the theoretical to the practical in just the last few years. But at present there are but few reported methods for understanding multiple locus systems and we do not know enough about their efficiency, statistical properties, or relative strengths and weaknesses to evaluate them for the task at hand. In addition, the difficulties of family-based association tests have led investigators to develop methods to overcome population stratification and even to question how much of a problem it represents. The focus of our work has always been on understanding how gene-gene interaction affects the analysis of genetic data. The current application is a plan for exploring the behavior of two-and-more locus models in human disease, an area in which our group has been active since 1981. We will emphasize detecting gene-gene interaction, as well as the power to detect multilocus systems. We will study how clinically variable phenotypes confound locus detection, investigate the limitations of heterogeneity analysis in linkage data, and examine the limits of the effect of population stratification as well as test several methods to correct for population stratification in case-control studies. The proposed work will examine four topics that are related to finding genes for common disease: (1) Determining gene interaction from linkage data; (2) Exploring the behavior of interactive multilocus models and our ability to identify those loci; (3) Understanding the widely-used admixture test for linkage heterogeneity; and (4) Exploring the effects, and corrections for, population stratification for qualitative and quantitative traits.
Showing the most recent 10 out of 95 publications