View article

Understanding of the Tree of Life has increased since genome-scale datasets became available. However, there are some nodes that remain unresolved. Examples of such 'tricky' nodes include the root of the animal tree or the internal relationships within Ecdysozoa. In this thesis, I investigate these two problematic nodes using state of the art and accurate computational methods. Firstly, I use recoding approaches to address the root of the animal tree. Amino acid recoding has been broadly used to reduce the inherent heterogeneity of genome-scale datasets, but its efficacy has been rarely tested empirically. Here I use simulations to show that recoding improves phylogenetic accuracy. Moreover, I show that results from real datasets addressing the root of the metazoan tree follow predictions from simulated data and suggest that phylum Porifera represents the sister-group of all the other animals. Then, I focus on the phylogenetic relationships among ecdysozoan phyla, studying in particular on the ambiguous phylogenetic position of phylum Tardigrada. I use dataset-specific site-heterogeneous substitution models to infer a phylogeny of Ecdysozoa, finding the alliance between tardigrades and onychophorans and arthropods as the most plausible evolutionary hypothesis for this meiofaunal group. Moreover, for the first time using sequence data, I show that the clade Nematoida might be an artifact. Finally, I investigate further the origin of ecdysozoans. The fossil record of this group is among the best understood of major animal clades and it is believed to document their evolutionary history well. However, molecular clock analyses have implied a …