The human molecular clock and mutation process : a characterization using microsatellite DNA

Abstract

In the past decade, thousands of human genomes have been catalogued, either by whole-genome sequencing or by targeted genotyping. The variability between human genomes encodes invaluable information about human traits and genetic diseases, as well as human migration patterns and population interactions. A key challenge is to understand and characterize the evolution of the variability between human genomes. In this thesis, I focus on studying human evolution through the use of microsatellites, which are simple repetitive sections of DNA of typically 1-6bp motifs (e.g. CACACACACA) that are highly polymorphic and highly mutable. The first aim is to establish that microsatellites are useful as reliable molecular clocks, such that its evolution highly correlates to time, especially when applied to the time range appropriate for human history. Using existing models of microsatellites, we examine microsatellite data from populations around the world to demonstrate that microsatellites are accurate molecular clocks for coalescent times of at least two million years. These results raise the prospect of using microsatellite data sets to determine parameters of population history. In order to calibrate genetic distances into time, the mutation rate must be known. This leads to the second aim, which is to directly measure the microsatellite mutation rate from largescale pedigree genetics data and provide a precision that is unprecedented. To do so, we use data from over 95,000 individuals in Icelandic pedigrees, genotyped in over 3000 microsatellite loci. Using trio and extended-family based approaches, we discover 2058 denovo mutations. In addition, we also attempt to capture many features that are covariates with the mutation rate, such as parental gender and age. The third aim takes our empirical observations of the microsatellite mutation process to build a new model of microsatellite evolution. This model improves upon the standard random walk model with features we have captured from aim 2. We use a Bayesian coalescent approach to provide a model that estimates the sequence mutation rate, European genetic divergence times, and human-chimpanzee speciation time.