Genomic Analysis of Human Population Structure
Recent developments in technology and computation have encouraged a shift towards a whole-genome approach to genetic analysis. Two key contributors to this shift, the Human Genome Project and the HapMap project, sparked an interest in studying the genetic patterns found in particular groups of individuals. The Maori population of New Zealand is an ideal, yet untapped, model for such studies due to recent partial mixture of two distinct population groups, and a culture of good documentation of genealogical information. A previous study carried out by the author found observable genetic differences between Maori and European populations in markers of forensic significance, yet no particular genetic patterns were found that were uniquely Maori. This study extends the previous work by developing methods to determine to what scale these differences exist, as well as demonstrating that a knowledge of these differences and methods could be used to improve current practices for clinical diagnosis. The current project began by taking a ‘candidate gene’ approach, studying two regions where there were known large genetic differences between Maori and European individuals: the region of Alcohol Dehydrogenase genes on Chromosome 4 (Chapter 2), and the Monoamine Oxidase A gene region on Chromosome X (Chapter 3). In both of these regions, large frequency differences were observed between Maori and non-Maori populations at both a single mutation level, and at a haplotype level. Despite the differences that were observed, no particular combinations of mutations could be considered uniquely Maori or uniquely non-Maori, so studies were expanded to the entire genome. This epansion was made possible due to the recent and continuing developments in genome-wide technology and advancements in computational speed and efficiency. Once it was possible to carry out a genome-wide study of genetic differences, the goal of research changed from determining whether or not Maori and European individuals were uniquely different at a genotype level, to how small a marker set could be produced while maintaining population-uniqueness at a genotype level. A method that uses bootstrap sub-sampling and other internal validation techniques has been developed for the generation of such a signature set for a Maori tribe (Ngati Rakaipaaka), and the generated set has been validated in other similar populations (Chapter 4). As a consequence of producing this set, the degree of European admixture was estimated in the tribe (28.7%), with over 15% of individuals within Rakaipaaka found to have no discernible European genomic ancestry. In a validation of the signature set generation method itself, the marker selection procedure was repeated for Type 1 Diabetes, a disease with high heritability. An analysis of case and control individuals using this signature set found that the generated set is able to perform better than a genome-wide reference set of mutations known to be associated with Type 1 Diabetes. This validation study, other potential uses, and a more detailed discussion of the signature set generation method are presented in Chapter 5.