Degree
Doctor of Philosophy (PhD)
Department
Biological Sciences
Document Type
Dissertation
Abstract
Phylogenetic inference has the ultimate goal of clarifying historical relationships between evolutionary units, and phylogenetic results can be used to make downstream inferences about the mechanisms of the evolutionary process. As genomic scale datasets have become more readily available over the past three decades, the statistical power with which we make these inferences has increased, but so has the variation in empirical data with which our models of evolution are forced to contend. In many cases, discordance can be observed between even strongly supported estimates of phylogenies, and while sometimes the sources of this discordance can be easily identified, other sources can be more cryptic. Despite some current methods smoothing over this discordance by arriving at a single phylogenetic estimate that attempts to account for all variation, the heterogeneity that underlies many phylogenetic datasets can often provide valuable insight into evolutionary processes and the methods with which we study them. In this dissertation, I explore the discordance and variation found in phylogenetic results in a variety of scenarios and examine the ways that different types of variation can be detected, identified, and used to increase our understanding of evolution. First, I explore gene tree discordance found in mitochondrial genomes first observed with simple models of sequence evolution and find that even complex models fail to increase concordance among inferred mitochondrial gene trees. Next, I investigate the phylogenetic placement of red pandas using a newly assembled phylogenomic dataset, and I investigate and identify the sources of discordance present in the dataset using a variety of methods and find evidence of both systematic and biological variation. Finally, I design and implement a new posterior predictive test statistic for assessing absolute model fit on a branch-specific basis and test its implementation in a simulated scenario. This is the first such test of its kind, and I use it to detect poor branch-specific model fit in an empirical dataset concerning the phylogenetic placements of Turtles and members of Archosauria.
Date
4-17-2024
Recommended Citation
Toups, Benjamin S., "Disentangling the Roots of Gene Tree Variation" (2024). LSU Doctoral Dissertations. 6428.
https://repository.lsu.edu/gradschool_dissertations/6428
Committee Chair
Brown, Jeremy M.