| For my Ph. D. dissertation, I have studied three subjects of molecular evolution by using statistical methods. (1) Evolutionary changes of the target sites of two microRNAs encoded in the Hox gene cluster of Drosophila and other insect species. (2) Patterns of duplication of microRNA genes and testing if the duplication patterns can support the hypothesis of genome duplication in the teleost fish lineage. (3) Statistical studies of the effects of random fluctuation of selection intensity on the extent of genetic variation. I chose these three subjects, because they are important and I could test hypotheses proposed with recently obtained empirical data. First, I examined the evolution of miRNA genes, MIR-iab-4, MIR-iab-4as, and MIR-10, which are encoded within the Hox gene cluster in invertebrates, and their target sites of 12 Drosophila species, other insects, and Daphnia. The results showed that the numbers and nucleotide sequences of these miRNA genes have been well conserved among the species examined, whereas the number of target sites has frequently changed during the evolution of insects and Daphnia. I also found that the old target sites, which were generated before the divergence of 12 Drosophila species, have been highly conserved through purifying selection. By contrast, some new target sites, which were generated after the divergence of 12 Drosophila species, have often been lost under relaxed selection. Second, I used miRNA genes to test the currently debated hypothesis that an ancestral species of teleost fishes experienced genome duplication. In this study I identified the miRNA genes that existed before the divergence of teleost fishes and tetrapods and examined their evolutionary changes in the teleost and tetrapod lineages. This study could support the hypothesis of genome duplication in the teleost fish. Finally, I conducted a mathematical study on the probability of fixation and fixation time of a new mutant allele and the extent of genetic variation expected under the currently competing models of fluctuating selection; competitive selection and stabilizing selection. As expected, the stabilizing selection model was shown to generate more genetic variation than the competitive selection model. However, when I analyzed the 10,602 polymorphic sites of SNP data from the human populations in Africa, the results obtained showed that the neutral model fit the data quite well and therefore the effect of fluctuation of selection coefficient is not very important whether the fluctuation occurs following the competitive or stabilizing selection models. More detailed results obtained for the three subjects studied are presented in the summaries of chapters 2, 3, and 4. The introduction for each chapter is also presented separately, because the subjects studied are somewhat disparate. |
