Dissertation/Thesis Abstract

The evolution of gene regulatory architecture: Insights from modeling natural selection and biological function
by Bullaughey, Kevin, Ph.D., The University of Chicago, 2012, 224; 3499713
Abstract (Summary)

Understanding the evolution of biological systems involves understanding (i) how mutational changes to the genetic encoding of the system alter system function (ii) how these functional changes impact fitness, and (iii) the population processes dictating the fates of these changes. These basic components of functional evolution are interdependent, although they are often studied individually. At the crux of these components is the extensive epistasis exhibited by biological systems; both the effect of a mutation on the traits comprising the system, as well as the fitness effect of altering the traits, may be highly dependent on other aspects of the system (i.e., the genetic background). Therefore, understanding the evolution of function requires understanding the epistasis intrinsic to the systems of interest, and the evolutionary implications of that epistasis. Fortunately, our understanding of the function and encoding of biological systems is reaching the point that realistic models of these can be combined with population genetic models in order to shed light on many outstanding questions in evolutionary biology. I pursue this approach of uniting functional and population genetic models to address several questions related to gene regulatory evolution.

The first question I examine is, how can the encoding of regulatory control systems evolve in the presence of efficient stabilizing selection that preserves functional output? I combine a mechanistically-motivated model of cis-regulatory function that has been shown to be predictive in Drosophila melanogaster, with a model of stabilizing selection on a multi-dimensional expression phenotype. I show that functional turnover is due in a large part to shifting patterns of constraint and that even the overall level of constraint evolves considerably. I also show that the substitutions most responsible for functional repatterning often would have been deleterious had they occurred on earlier haplotypic backgrounds. This study highlights the functional importance of nearly-neutral mutations and reveals a high degree of historical contingency in the evolution of regulatory DNA.

In the second study, I examine whether it is plausible that substitutions of opposing functional effect can be individually adaptive given the epistasis likely to characterize regulatory networks. As a case study, I employ a mathematical model of a three-gene network, the type I coherent feed-forward loop, which is ubiquitous in nature, accurately described by the model, and has an interpretable basis for fitness. I show that during a single bout of adaptation, substitutions that affect a trait in one direction can be beneficial while subsequent substitutions that affect the trait in the opposite direction can also be beneficial. The existence and apparent prevalence of such adaptive reversals has important implications for interpreting comparative genomics data and for detecting adaptation.

These two studies, which comprise the main part of my dissertation, exemplify the utility of combining functional and population genetic models to study evolution. In particular, the epistasis present in these model systems stems directly from the functional properties of the models, rather than an arbitrary specification of epistatic effects (as is often modeled). In turn, the evolutionary dynamics are heavily influenced by the epistasis, and the modeling yields new, detailed insights regarding functional evolution.

Indexing (document details)
Advisor: Przeworski, Molly
Commitee: Borevitz, Justin O., Hudson, Richard R., Kreitman, Martin, Ruvinsky, Ilya
School: The University of Chicago
Department: Ecology and Evolution
School Location: United States -- Illinois
Source: DAI-B 73/07(E), Dissertation Abstracts International
Subjects: Genetics, Evolution and Development, Bioinformatics
Keywords: Computational biology, Gene expression, Gene regulation, Natural selection, Population genetics
Publication Number: 3499713
ISBN: 9781267246936