Uncovering and understanding the processes that have led to the biological diversity we observe today are of fundamental interest in biology. Since direct observation of speciation is usually impossible, knowledge about the processes behind species formation can be gathered by studying mutations, natural/sexual selection, and genetic drift. In this thesis I aim to identify evolutionary processes that cause species divergence and, ultimately, speciation using Swedish cyprinid fish as a model system. Assuming that the demographic history of a population is mirrored in the genome, I studied the effects of a bottleneck on genetic variability in populations of roach. As expected, I found that a decrease in population size caused a decrease in genetic variability, a pattern that was obtained from both microsatellite and mitochondrial data. The importance of hybridization for speciation is debated, however, by analyzing morphology and microsatellites I could show that common bream and white bream and their interspecific hybrids are phenotypically and genetically differentiated and that ongoing geneflow is mainly unidirectional. Ongoing geneflow antagonizes the effect of genetic drift, but by studying isolated populations (= no gene flow) the impact of genetic drift can be assessed. Long-term isolated populations of roach and perch surprisingly showed stable levels of genetic diversity over time despite decreasing effective population size. However, each population genetically diverged during the period of investigation, a finding that is consistent with the effect of drift. An analysis of the systematic relationship of the 18 species of Swedish cyprinids revealed low congruence of phylogenies based on two different genetic markers. The position of the tench remains unresolved and the relationship of common bream and white bream as sister species cannot be confirmed. Within cyprinid fishes, diversification rates reveal a slowdown with time, a pattern that I found also in other fish clades and that is consistent with density-dependent cladogenesis. Overall, based on the findings presented in this thesis I emphasize that the maintenance of genetic variation in populations is essential since genetic variation is the key element for processes of divergence to act upon.

The research presented in this dissertation integrates data and theory to examine three important topics in biology. In the first chapter, I investigate genetic variation at two loci involved in a genetic incompatibility in the genus Xiphophorus. In this genus, hybrids develop a fatal melanoma due to the interaction of an oncogene and its repressor. Using the genetic variation data from each locus, I fit evolutionary models to test for coevolution between the oncogene and the repressor. The results of this study suggest that the evolutionary trajectory of a microsatellite element in the proximal promoter of the repressor locus is affected by the presence of the oncogene. This study significantly advances our understanding of how loci involved in both a genetic incompatibility and a genetically determined cancer evolve. Chapter two addresses the role polyploidy, or whole genome duplication, has played in generating flowering plant diversity. The question of whether polyploidy events facilitate diversification has received considerable attention among plant and evolutionary biologists. To address this question, I estimated the speciation and genome duplication rates for 60 genera of flowering plants. The results suggest that diploids, as opposed to polyploids, generate more species diversity. This study represents the broadest comparative analysis to date of the effect of polyploidy on flowering plant diversity. In the final chapter, I develop a computational method for designing disease surveillance networks. The method is a data-driven, geographic optimization of surveillance sites. Networks constructed using this method are predicted to significantly outperform existing networks, in terms of information quality, efficiency, and robustness. This work involved the coordinated efforts of researchers in biology, epidemiology, and operations research with public health decision makers. Together, the results of this dissertation demonstrate the utility of applying quantitative theory and statistical methods to data in order to address complex, biological processes. / text