The objective of this project is to develop statistical methods and markers for mapping genes that underlie ethnic differences in disease risk, based on a novel approach that exploits the autocorrelation of ancestry on chromosomes of mixed descent in a manner analogous to linkage analysis of an experimental cross. This has obvious applications to investigating the genetic basis of conditions such as hypertension, Type 2 diabetes and systemic lupus erythematosus in the USA and other countries where there has been recent admixture between ethnic groups that have different risks of disease for genetic reasons. The development of the statistical analysis program will rely on methods developed for """"""""missing data"""""""" problems, using Markov chain simulation to generate the posterior distribution of ancestry at each locus given the observed marker data at all loci. A score test for linkage will be obtained by averaging over this posterior distribution. Extensive tuning of the algorithms will be required to ensure that the the statistical methods are robust over all models and data structures that are likely to arise in practice. The development of the marker sets will extend the work already in progress in Shriver's lab, based on screening SNP libraries and using subtractive hybridization to identify population-specific alleles. As the marker set is developed, the multipoint analysis program will be used to plot the information content for ancestry extracted by the marker set and to identify areas where additional markers are required to meet the target of extracting 80 percent information about ancestry in an initial genome searches. The program and the markers will be beta-tested on data from studies in admixed populations now under way in the USA, the Caribbean, and Australia The program will be made publicly available via the web.
