Browsing by Author "Kerminen, Sini"

Studies of population structure are motivated by the need to understand population history and to have well-characterised groups of individuals in studies of genetics of diseases and traits. A standard method to analyse genetic population structure is principal component analysis (PCA). A disadvantage of PCA is that it can reliably handle only independent genetic markers. This means that the genetic markers that are correlated with other genetic markers have to be excluded from the data. This leads to a loss of information. In 2012, Lawson et al. published a chromosome painting method that can utilise haplotype information, i.e. information from correlated markers, and thus it can detect more subtle differences in populations than the standard PCA. This thesis studies two questions. The first question is whether the chromosome painting method can provide more precise genetic clustering of geographically defined Finnish groups than the standard PCA method. The second question is whether the chromosome painting method can reveal new details of population structure in Finland. The data used in this study are from the FINRISK Study survey of 1997. This cohort includes the genotype data of about 4,000 individuals and the information about individuals and their parents birthplaces. 345 Individuals were randomly chosen from the cohort in such a way that both of their parents were originated from the same province. Ten provinces of Finland were used as study groups for the method comparison. First, the data were analysed with SmartPCA (a standard PCA method) and ChromoPainter (the chromosome painting method) and the results were compared both visually and quantitatively. Finally, the individuals were assigned to populations based on the ChromoPainter result using FineSTRUCTURE program and these genetic populations were compared to the geographic origin of the individuals. The results showed that the chromosome painting method clustered seven out of ten groups significantly tighter than the standard PCA. Nevertheless, SmartPCA was faster and easier to use than ChromoPainter. The main population genetic division was found between the eastern and western parts of Finland, which was consistent with earlier studies. All in all, 15 populations were detected and the results revealed that they were geographically clustered. The genetic populations correlated well with the borders of Finnish provinces and counties. As the first conclusion, the chromosome painting method was able to give more precise results than the standard PCA but the standard PCA is still more suitable for quick preliminary analyses of genetic data. As the second conclusion, the chromosome painting method was able to detect detailed subpopulation structure in Finland and these populations are geographically clustered. Results provide an excellent basis for the future studies of population structure and genetic diseases in Finland.