Understanding predictability of evolutionary trajectories

Work over the last decade has provided tantalizing clues that the genetic mechanisms that underlie evolutionary trajectories may be more constrained and deterministic than previously appreciated. Whether the genetic basis of evolution is predictable or stochastic is a fundamental question in evolutionary biology, and the ability to predict evolutionary responses has critical implications for the

fields of agriculture, medicine and conservation. Unfortunately, we are far from understanding the factors that underlie the predictability of evolutionary responses or the fitness consequences associated with alternative evolutionary trajectories in nature. Surveys of replicate populations adapting to the same environmental conditions provide an opportunity to quantify the repeatability of

evolutionary trajectories and determine the factors affecting observed patterns of variation. Several genetic factors are hypothesized to affect the probability of a gene or mutation’s use during the process of adaptation. However, to date there have been few empirical tests determining the contribution of individual factors to patterns of gene re-use (parallelism) in natural populations.

In the proposed work, we built upon our prior findings and combine innovative population genomics, genetic engineering, and experiments to answer three core biological questions: 1) How does parallelism translate across biological levels? 2) Which genetic factors affect parallelism? 3) What are the fitness consequences associated with genetically distinct, yet

phenotypically similar, evolutionary trajectories? Our laboratory is well positioned to answer these profoundly important questions about the repeatability and predictability of evolutionary trajectories. Our model system for addressing these questions is the threespine stickleback (Gasterosteus aculeatus), a species with bountiful genetic and genomic resources. Stickleback are

considered a textbook example of parallel phenotypic evolution; thousands of independently derived stickleback populations have repeatedly adapted to a variety of freshwater environments. We will survey several focal freshwater populations to examine the scaling of papalism across biological levels and to test whether the source and type of genetic variation or epistatic interactions predict the

frequency of gene use during adaptation. The creation of genetically modified lines that differ only in their allele frequencies at candidate loci will allow us to test whether fitness coefficients can predict the probability of gene use (magnitude of genetic parallelism) in nature. Together this work will test

several fundamental evolutionary questions and take a first step toward determining whether the development of a predictive evolutionary framework is possible.