Most human diseases exhibit some level of familial clustering, indicating the role of inherited factors. But in most cases, the mode of transmission does not fit simple Mendelian models, and disease risk is probably determined by non-additive interactions between multiple genetic and environmental factors. One of the central challenges in current human genetics is to unravel the basis of these so-called "complex" diseases. The on-going maturation of genomic tools (including genome sequences for humans and other organisms, SNP databases, and the planned haplotype map), as well as advances in genotyping technology, will make it significantly easier to identify genetic variants that contribute to complex disease. However, the central problem in complex disease mapping is that we aim to find genetic variants that, on their own, make only a relatively small contribution to total disease risk. This fact indicates the importance of sound statistical methods that can efficiently extract weak signals from noisy data. In this proposal, we outline three closely related projects connected with linkage disequilibrium-based mapping. (1) We will develop a novel multipoint method for both association mapping and LD fine mapping that we anticipate will be more powerful and more robust than existing methods. (2) We will examine publicly available data sets of human linkage disequilbrium in order to further characterize the biological basis of patterns of LD, and to get a deeper understanding of the nature, and extent of the recently proposed "haplotype blocks." (3) We will examine various critical questions of experimental design, including how to choose the optimal SNPs to map a region and how to make the most efficient use of the forthcoming "haplotype map." The long term goal of this project is to develop statistical/computational tools and models of LD that can be of practical use to complex trait mapping community, and will contribute to cheaper and more effective studies.

R01HG002772

2003-04-10

2009-03-31