Genetic Draft and Linked Selection in Rapidly Adapting Populations

Abstract

Evolution proceeds by the stochastic emergence and fixation of mutations over time. The dynamics of this process are well understood if beneficial mutations are sufficiently rare or the population is sufficiently small. However, as the population size increases or beneficial mutations become more common, many selected mutations may arise and spread on different genetic backgrounds, competing with each other for fixation. In this case, the fates of alleles depend not only on their direct effect on fitness, but also on the genetic backgrounds on which they are found, and on their associations with future mutations that may arise in the population. The stochastic fluctuations in evolutionary dynamics that result from these factors are referred to as genetic draft, in reference to the other primary stochastic force in evolution, genetic drift. In this thesis, I investigate how genetic draft alters the trajectories of neutral and selected alleles and skews patterns of genetic diversity, with an emphasis on tying theoretical predictions of evolutionary dynamics to experimental inference and observations. The first part of this thesis investigates the dynamics of mutant alleles from a theoretical perspective that includes the evolutionary forces of mutation, selection, and drift. In this section, I analyze a simple model of genetic draft in rapidly adapting, asexual populations. I derive a prediction for how the frequencies of selected and neutral alleles will change over a characteristic timescale, which I use to infer the typical fixation and sojourn times of selected mutations. I also derive several statistics of genetic diversity, including the average heterozygosity and site frequency spectra. Finally, I use such predictions to analyze how genetic draft skews commonly used tests for selection and propose a modified test that accounts for this additional stochasticity. The latter half of the thesis explores experimentally how genetic draft affects the speed, diversity, and genetic basis of adaptation at different recombination rates. I evolve hybrid populations of budding yeast with different rates of outcrossing, while tracking genetic diversity, the rate of adaptation and the variance in fitness over time. I observe experimentally a phase transition from highly deterministic and polygenic adaptation at high recombination rates to the stochastic emergence and sweep of clones as the rate of outcrossing decreases. I investigate the consequence of this transition for the speed and predictability of adaptation. Finally, by comparing experimental observations with simulations, I infer that adaptation at high recombination rates is driven by a large number of dense, weakly selected variation, and examine how linkage and interference between these variants distorts the trajectories and outcomes of adaptation from simple population genetic predictions.