Statistical analyses of genealogical-phylogenetic data

Thanks to the recent advancement of the sequence technologies, generating large volumes of DNA sequence data is now becoming more feasible. Sequencing several samples across many species from a range of clades enables us to connect the two fields of study previously separated due to the lack of data: population genetics and phylogenetics. The former has focused on detailed genetic processes in a few species, while the latter has studied large-scale evolutionary relationships across many species. In this thesis, methods to utilize the new type of data, genealogical-phylogenetic data, are explored to tackle the problems lying between the two fields, including how to delimit species with genetic information and how ecological traits affect species genetic properties. First, a method of species delimitation based on single locus gene tree, called the generalized mixed Yule coalescent method (GMYC method), is evaluated. Its statistical properties are assessed on both simulated and real data, and the method is extended to relax some simplifying assumptions and to give a robust confidence measure. The simulation studies showed that the reliability of the delimitation depends on population parameters and patterns of diversification processes. Assessment of the performance on a dataset of 5196 water beetle mitochondrial DNA sequences sampled from across Europe showed that the method accurately delimited half of the studied species. The accuracy was affected by several factors, notably the presence of pseudogenes and potential undersampling of species range. Then, the water beetle data and the GMYC method are used to test the effects of species ecological traits on genetic properties, focusing on species habitat type. Habitat type had significant effects on genetic variation and substitution rate via effects on range size and latitudinal distribution of species. However, direct effects of habitat type on genetic properties were not observed.EThOS - Electronic Theses Online ServiceGBUnited Kingdo