Molecular Evolution and Systematics of New Zealand Cicadas

Abstract

This thesis contains results from the phylogenetic analysis of DNA sequence data from a range of New Zealand, Australian and New Caledonian cicada species. The aim of this research was two-fold. First, the inferred phylogenetic relationships among populations, species and genera are used to evaluate previous biogeographic and evolutionary hypotheses relating specifically to the New Zealand cicada fauna and extended to the entire New Zealand biota in general. Second, the nucleotide sequences were used to further examine the effects of biological assumptions on the alignment of DNA sequences and the estimation of phylogenetic relationships from these alignments. Three, partially overlapping studies were carried out in order to accomplish the primary aim. The first of these concentrates on the phylogeographic relationships among populations of the widespread species Maoricicada campbelli using mitochondrial DNA sequence data. This study has revealed a deep and previously unsuspected phylogenetic split within the species. This split is explained as the product of historic environmental change, in particular, rifting along geological faults and other associated tectonic changes. A second study, on the phylogenetic relationships among species of the genus Maoricicada, has shed light on the origin and evolution of the New Zealand alpine biota. In addition to conventional tree topology reconstruction, I have implemented the recently developed Multiple Comparison Test for discriminating between multiple alternative phylogenetic hypotheses. These analyses have revealed that the radiation of alpine Maoricicada species taxa may have involved an evolutionary reversal, whereby montane progenitors have given rise to two low altitude species, a pattern not suspected by previous workers. However, the alternative interpretation remains that the alpine species are monophyletic and are nested within a radiation of low altitude species. Third, a much larger data set of mitochondrial and nuclear coding DNA sequences were used to infer the relationships among the five New Zealand cicada genera and putative Australian and New Caledonian outgroups. I have implemented two likelihood-based tests; the parametric bootstrap and the widely used Kishino-Hasegawa test to detect significant incongruence between estimates of topology inferred from multiple mitochondrial genes and a single nuclear gene (EF1α). These analyses revealed significant incongruence in phylogenetic signal between the mitochondrial genes and the nuclear EF1α gene. I used heterogeneous maximum likelihood models and non-parametric bootstrapping to attempt to resolve between the possibilities of systematic error and separate histories as the cause of this incongruence. The second aim of my thesis involved two studies. The first of these was an examination of the role of nucleotide substitution models, with an emphasis on among-site rate variation models, on phylogenetic analysis. It is shown that selection of an appropriate substitution model is essential for the correct selection of an optimal tree topology and the estimation of edge lengths and substitution model parameters. I have also shown that changes in the assumptions in applying substitution models can have a dramatic effect on statistical support for phylogenetic hypotheses, as inferred by non-parametric bootstrapping. I also identify a possible example of the "anti-Felsenstein zone", where maximum likelihood and minimum evolution are behaving in an inaccurate manner with regard to estimation of topology, and argue that this problem can be alleviated by using a more realistic substitution model. The second study into the role of biological assumptions in phylogenetics is the presentation of a refined secondary structure model for domains IV and V of the insect mitochondrial large subunit rRNA gene. Conserved motifs and secondary structural elements are described, based on the comparative analysis of over 400 insect sequences from twelve different orders. Development and refinement of such structural models is essential if an alignment of DNA sequences is to reflect homology and not similarity. These motifs and conserved secondary structural elements will aid in the alignment in as yet unsequenced insect taxa and the drawing of secondary structure models.