Linkage and segregation analysis are two key tools of the genetic epidemiologist. Improved design and analysis of family studies of linkage and segregation are the goals of this proposal. I address these goals by proposing to solve a set of methodological problems. First, I will develop methods to evaluate the statistical power of proposed genetic linkage studies. I will deal with complex genetic models and possible genetic heterogeneity, as is required for diseases such as schizophrenia and diabetes. Since multipoint linkage provides greater power than two-point linkage, I will also consider power calculations for multipoint linkage. Development of methods to evaluate the power of a linkage study prior to its execution will represent a timely contribution, because of current advances in molecular genetics and the rapid rate at which new linkage studies are being undertaken. Second, I will compare two current methods and one new one for constructing a confidence interval for the recombination fraction. Such estimates are used in genetic counseling for Mendelian disorders such as Huntington disease. Third, I will examine a likelihood-based strategy for detecting families segregating at a major locus for a quantitative trait. Because quantitative physiological traits are often risk factors for complex diseases, this strategy has the potential to identify subgroups of individuals with different etiological bases for disease. It has the further potential to identify families on which subsequent linkage analysis will be most fruitful. Fourth, I will derive methods for incorporating prior information about genetic model parameters into family studies of genetic segregation. Such prior information might consist of the results of a previous cross-sectional study. The value of such an approach would be considerable because of the high cost of gathering family data and the frequent availability of prior information. Moreover, the project will not necessarily be restricted to the specific areas detailed above. As time permits during the grant period, and as particularly interesting problems arise, I will also devote time to them.
