Novel Phylogenomic Methods for Uncovering the Evolutionary History of the Hyperdiverse Clade Caryophyllales

Abstract

Gene tree conflict has historically been attributed to methodological error, but can also arise from an array of biological processes (e.g. hybridization, incomplete lineage sorting, etc…). Affordable genome sequencing has made it possible to study patterns of gene tree conflict, which in turn provides evidence for the evolutionary events that led to this conflict. This thesis applies and formulates novel methods for analyzing gene tree conflict to better understand the evolutionary history of the hyperdiverse plant clade Caryophyllales. The uniparentally inherited chloroplast genome is used as a basis for understanding the degree of gene tree conflict that arises from systematic error. Despite typically acting as a single locus, whereby all genes share one evolutionary history, I examined a large chloroplast genome dataset that showed substantial gene tree conflict. Conflict was significantly correlated with the length of the genes’ alignment and the amount of evolutionary information within the gene. I also examined sources of gene tree conflict with a transcriptome dataset of the carnivorous clade within the Caryophyllales. The carnivorous clade contained evidence of ancient hybridization and high levels of gene/genome duplications. Every family in the clade, except one, had a unique paleopolyploidy event. The most highly debated node of the carnivorous Caryophyllales had a large amount of gene tree conflict, and phylogenetic results were highly sensitive to taxon sampling. Using the largest phylogenomic dataset to date, I also reconstructed the evolutionary relationships among lineages across the entirety of the clade Caryophyllales. This was done using a new computationally tractable and phylogenetically informed method of hierarchical clustering. The results provide an outline of highly conflicting regions in the clade Caryophyllales and provide evidence for a new family, Agdestidaceae, which is proposed based on morphology and phylogenetic position. Furthermore, despite the large amount of data, the inferred phylogeny was highly congruent with the three gene phylogenies found in the literature. This demonstrates that our current hypotheses of the tree of life are reassuringly congruent with larger datasets. The limited ways the data could be analyzed emphasizes the need for advanced methods that utilize gene tree conflict to inform species relationships. To fully embrace the information provided by gene tree conflict, this thesis presents a new method of analysis that calculates the likelihood for a single relationship on the phylogenetic tree. By isolating a single relationship of interest, the other relationships are allowed to vary, which provides a means of incorporating the gene tree conflict into the analysis of the relationship. This method is used on the contentious relationship in the carnivorous clade and further validated using a vertebrate dataset. The results show that while the phylogenetic reconstruction supports one hypothesis, the data itself supports an alternate hypothesis. This presents a novel framework to conduct analyses that are focused on a single contentious relationship at a time. Overall, this thesis results in the development of novel methods for phylogenomic analyses to provide evolutionary insights into the hyperdiverse Caryophyllales.