Return to search

Discrete and statistical approaches to genetics

This thesis presents a number of major innovations in related but different areas of research. The contributions range along a continuum from mathematical phylogenetics, to development of statistical methodology for detecting recombination and finally to the application of statistical techniques to understand Feline Immunodeficiency Virus (FIV) an important pathogen. An underlying theme is the application of combinatorial and statistical ideas to problems in evolutionary biology and genetics. / Chapter 2 and Chapter 3 give a number of results relevant to mathematical phylogenetics, in particular maximum parsimony. Chapter 2 presents a new formulation of maximum parsimony in terms of character subdivision, providing a direct link with the character compatibility problem, also known as the perfect phylogeny problem. Specialization of this result to two characters gives a simple formula based on the intersection graph for calculating the parsimony score for a, pair of characters. Chapter 3 further explores maximum parsimony. In particular, it is shown that a maximum parsimony tree for a sequence of characters minimizes a subtree-prune and regraft (SPR) distance to the sets of trees on which each character is convex. Similar connections are also drawn between the Robinson-Foulds distance and a new variant of Dollo parsimony. / Chapter 4 presents an application of the work in Chapters 2 and 3 to develop a statistical test for detecting recombination. An extensive coalescent based simulation study shows that this new test is both robust and powerful in a variety of different circumstances compared to a number of current methods. In fact, a simple model of mutation rate correlation is shown to mislead a number of competing tests, causing recombination to be falsely inferred. Analysis of empirical data sets confirm that the new test is one of the best approaches to distinguish recurrent mutation from recombination. / Finally, Chapter 5 uses the test developed in Chapter 4 to localize recombinant breakpoints in 14 genomic strains of FIV taken from a wild population of cougars. Based on the technique, three recombinant strains of FIV are identified. Previous studies have focused on the epidemiology and population structure of the virus and this study shows that recombination has also played an important role in the evolution of FIV.

Identiferoai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.102964
Date January 2006
CreatorsBruen, Trevor Cormac Vincent.
PublisherMcGill University
Source SetsLibrary and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada
LanguageEnglish
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Formatapplication/pdf
CoverageDoctor of Philosophy (School of Computer Science.)
Rights© Trevor Cormac Vincent Bruen, 2006
Relationalephsysno: 002590016, proquestno: AAINR32159, Theses scanned by UMI/ProQuest.

Page generated in 0.0022 seconds