Techniques for analysing the genetic epidemiology of complex diseases will be developed. Heart disease is a trait that clearly has both environmental and genetic components. However, understanding of how the genetic components contribute to disease risk has elluded investigators, who have used techniques designed for genetically simpler traits. The combination of current computer technology and novel simulation approaches to statistical estimation make timely the development of Monte Carlo methods of likelihood analysis. The Gibbs sampler is one such simulation technique, and can be applied on pedigrees, to produce samples from the underlying distribution of genotypic effects given phenotypic observations. These samples can be used to evaluate and maximise likelihoods for complex genetic models involving several heritable effects, environmental effects, and linked markers.

The proposed research is for the development and implementation of Gibbs sampler techniques in several areas area of genetic epidemiology. First, problems concerning multilocus Mendelian haplotypes, and multiple polygenic effects will be addressed. These methods will then be combined to develop methodology directed towards jointly performing linkage and segregation analysis for complex traits on extended pedigrees, and will include methods for estimating the parameters of complex genetic models. These techniques will be evaluated by analyses of a data set of pedigrees which have been collected for a study of environmental and genetic risk factors for coronary heart disease.