Whole-genome association mapping, with all its theoretical power to detect genetic variants that contribute to common disease, is finally becoming practical. Most methods for analyzing data from these studies have envisioned scans with hundreds of thousands of SNPs in a relatively homogeneous population such as European Americans. However, the differences that exist among human populations also need to be taken into account. Even in a population that is relatively homogeneous, cases and controls may have different ancestral histories, which will result in """"""""population stratification"""""""", or the population may be recently """"""""admixed"""""""" as is the case for African-Americans and Hispanics. We propose to develop tools & methods for Population Substructure Analysis (PSSA) to deal with these issues in a disease-mapping scenario. ? ? (1) Our first aim will be to improve our already published methods and software (ANCESTRYMAP) for admixture mapping. Admixture mapping is a method for carrying out a genome-wide association study in a population of recent mixed ancestry such as African or Hispanic Americans, with far fewer markers than are needed for a homogeneous population. In the past two years great strides have been made in turning admixture mapping into a practical method, and we expect to continue to extend its applicability. ? ? (2) Our second aim will address the problem that whole-genome association scans with hundreds of thousands of SNPs will be severely compromised in their power to study a minority population such as African or Hispanic Americans unless methods are developed that search for association after inferring an individual's ancestry state at each point in the genome. A key aim of PSSA is to build methods that allow fully-powered whole-genome association scans in minority groups. ? ? (3) Our third aim will be to provide a novel approach for correcting of population stratification in whole-genome association scans. Population stratification refers to systematic differences in ancestry between cases and controls, which can lead to allele frequency differences and false-positive associations. Building on previous work we introduce new methods to measure and correct for stratification. We believe our new techniques will provide near-optimal power, and will be computationally efficient. We intend to make all these tools publicly available for the scientific community. ? ? ? ?
Showing the most recent 10 out of 20 publications
