Doctor of Philosophy (PhD)
Identifying the source and structure of variation in nature is crucial to understanding fundamental aspects of evolution. Despite a recent plethora of genetic and morphological data, many interesting questions about the relationships between different groups remain unresolved. My dissertation evaluates three approaches for identifying and quantifying the variation within phylogenetic datasets. Characterizing variation within datasets and across analytical methods gives insight into biologically interesting characters, unusual evolutionary processes, and areas for model improvement.
Network-based community detection approaches offer a powerful tool to describe variation in phylogenetic signal across the genome (i.e., gene tree variation). In Chapter 2, I investigate the performance of recently developed network-based community detection approaches for identifying structure in phylogenetic data. Different network types excel at identifying clusters overlapping in tree space, clusters containing a drastically unequal number of trees, and efficiently analyze large datasets. Community detection methods provide a more quantitative approach to identifying phylogenetic structure that can be used on genome scale datasets.
In Chapter 3, I compare consistency across locus-by-locus support metrics using genome-wide sequence data from squamates to interrogate the unresolved placement of Iguania within the clade Toxicofera. I use marginal likelihood ratios and maximum-likelihood ratios to detect the direction and strength of phylogenetic support for a priori topological hypotheses. Despite significant correlation between different likelihood values, this correlation is driven by the tails of the distribution. Despite strong support placing Iguania somewhere within Toxicofera, there is not strong support for an exact placement.
Further exploring squamate phylogenetics, in Chapter 4 I use a topological mixture model to detect conflict across morphological characters. Using patterns of conflict, I assess the independence or modularity of subsets of characters. The majority of characters support the traditional placement of Iguania as sister to all other squamates. I find conflict within potential subsets, indicating a lack of modularity. Through these three chapters I compare methods and data sets to interrogate large phylogenetic datasets with the goal of identifying best practices for identifying conflicting information.
Mount, Genevieve Geraldine, "Quantifying Structure and Variation in Complex Phylogenetic Data" (2020). LSU Doctoral Dissertations. 5415.
Available for download on Saturday, October 28, 2023