EVOLUTIONARY IMPLICATIONS OF THE STANDING GENETIC VARIATION IN NATURAL POPULATIONS OF MIMULUS GUTTATUS

Abstract

The standing genetic variation present in a population is the raw material on which natural selection can act. The study of the architecture of this variation indicates how evolution might proceed. In this dissertation, I perform a detailed survey of this architecture both within and across natural populations of Mimulus guttatus. In the first two chapters, I focus on a single, natural population to demonstrate the role that interactions between genes (epistasis) affecting floral and developmental traits can have on the response to selection. More specifically, the first chapter takes greenhouse estimates of floral and developmental epistasis and demonstrates their contribution to the components of genetic variation. Contrary to popular belief, we find that epistasis largely determines the additive genetic variance in these traits, which directly controls the ability of a population to respond to selection. Moreover, this work suggests that the rate of change at individual loci and the ultimate fate of fixation depends on the frequency of alleles at all other interacting loci. In the second chapter, I demonstrate the natural relevance of the preceding chapter by showing that the epistasis observed in the greenhouse generates epistasis in fitness in a natural setting. In the final chapter, I develop a method to survey the segregating variation in flowering time across multiple populations. The minimal evidence of commonality in genetic architecture even among closely neighboring populations is a possible implication of epistasis (in addition other phenomena such as Genotype-by-Environment interaction). Overall, these observations of the standing genetic variation in natural populations imply a high degree of idiosyncrasy of evolution, in that different populations possess a unique set of variants with which to respond to selection, and even if variants were entirely shared, the genetic trajectory of evolution would depend on the unique set of starting allele frequencies in each population.