The purpose of this project is to develop methodology for analyzing molecular population genetics data. Current methods for localizing a disease gene using marker/disease linkage disequilibrium data do not adequately quantify how close th marker and disease locus are. We have developed a model based approach for estimating the recombination fraction between an RFLP marker and the disease locus, which unlike other methods provides confidence statements about the estimate. Th results show that the extra variability caused by the stochastic forces of evolution is not negligible and must be accounted for in the confidence calculations. Two extension of the model based approach have been considered. Th model has been extended to describe the joint behavior of two RFLP's and a method has been proposed to assess the likely topological relationship between the disease locus and two markers. Information of this type is useful for localizing the disease gene. The method was tested on cystic fibrosis data and gave correct topologies. A method was developed to estimate the mutation rate of a microsatellite marker which also provides confidence bound. If mutation is too large, then these polymorphic markers are not informative for marker/disease associations. A simulation study was carried-out to identify appropriate mutation rates.
