Methods for inferring human evolutionary history using genetic markers

This thesis includes a series of investigations into the problem of using neutral genetic markers to learn about population histories, and population structure. Patterns of genetic variation within species can convey a great deal of information about the history and ecology of that species, and about the nature of molecular evolutionary processes. An important property of population genetic variation is that genetic data from different individuals cannot be treated as independent: this means that specialized stochastic models are needed for population genetic analysis. Most of the work here relies heavily on the coalescent approach, which provides powerful techniques for statistical analysis of population data.; The five chapters in this volume include (1) an examination of the distribution of microsatellite variation in populations, under a range of demographic models; (2) a statistical test of heterogeneity of microsatellite variation across loci, and populations, designed to detect selection at linked loci; (3) a method of estimating, and placing confidence bounds on the ages of mutations, using linked variation; (4) a study of human Y chromosome variation, estimating the most recent common ancestor times of human populations, and estimating long-term demographic parameters of humans; and (5) a set of methods for analyzing case-control data for association gene mapping, in the presence of an unknown degree of population stratification.