On epistasis and adaptation in the budding yeast Saccharomyces cerevisiae

Abstract

The diverse forms of life which surround us today are the product of long lines of mutations, all running back to a common ancestor. Along these interconnected lines of evolutionary history, DNA sequences have been duplicated, deleted, mutated, and merged together by recombination, all while natural selection has steadily worked to sift through this variation and produce organisms that thrive in their environments. To understand this process, we must understand the fuel for evolutionary change: mutations. Recent advances in genetics, robotics, and sequencing have made it possible to measure the fitness effects of thousands of mutations at once, opening the door to large­-scale studies on the interactions between mutations, or epistasis. A clear result has emerged from this work: we often cannot predict the effect of a combination of mutations based on the effects of each mutation alone. This presents a problem for both understanding and predicting evolution, because the dynamics of adaptation depend directly on the fitness effects of mutations. In this thesis, I use open­ended experimental evolution and directed mutagenesis to characterize the predictability of epistasis and evolution in the budding yeast Saccharomyces cerevisiae. In Chapter 1, I measure the fitness effects of sets of insertion mutations in a panel of 163 offspring from a yeast cross, demonstrating that epistasis is widespread and that many mutations are more deleterious in higher fitness genetic backgrounds. In Chapter 2, I investigate the predictability of evolution by measuring phenotypic and genotypic changes in replicate S. cerevisiae populations over the course of a 10,000 generation evolution experiment. In Chapter 3, I measure how the fitness effects of mutations change in these evolving populations, showing how epistatic interactions can cause populations to become less robust to deleterious mutations as they adapt to their environment.