Date of Award

Spring 1-1-2019

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Nolan C. Kane

Second Advisor

Nancy C. Emery

Third Advisor

Rebecca J. Safran

Fourth Advisor

Samuel M. Flaxman

Fifth Advisor

John K. McKay

Abstract

This dissertation concerns the roles of genetic and environmental factors in producing trait variation in evolving populations, with an emphasis on the creation and use of statistical tools that facilitate predictions. The research concerns evolution across a variety of spatial and temporal scales and environmental conditions. In each study I employ statistical approaches to make predictions about how observed trait variation is derived from variation due to the environment, or genetics, or the interaction between the two. The first chapter investigates the evolution of species' performance curves through the construction of a Bayesian model that facilitates comparisons among groups. The model is used to investigate how performance curves have evolved among taxa in the genus Lasthenia, and how variation in their performance curves predict where they occur in nature with respect to fine-scale hydrological gradients. I find evidence that the microhabitats taxa occupy along fine-scale hydrological gradients is best predicted by their overall productivity rather than the conditions that optimize their performance. The second chapter concerns predictability of evolution during colonization. Using flour beetle microcosm experiments, this work demonstrates that the genetic and phenotypic predictability of evolution during colonization decays over time, highlighting the relative contributions of stochastic and deterministic forces that shape variation in dispersal, fecundity, and body size. The last chapter addresses key challenges in predicting phenotypes from genetic sequence data. I develop a novel approach to representing DNA sequences, and demonstrate its value to capture multiple types of genetic variation, which I then show can be effectively used as input for models predicting phenotype. Collectively, the three studies provide insights into the complex and interacting roles of genetic and environmental variation in generating traits, and the development and use of statistical methods to make predictions that are important to our understanding of evolutionary processes.

Share

COinS