Background: When designing genetic studies for any isolated human group it is crucial to understand it's population structure (i.e. founder effects, admixture, inbreeding) in order to fully harness the power of the target population in the detection of disease related genetic markers. The aim of this study was to characterise underlying patterns of genomic homozygosity in the Norfolk Island genetic isolate with the idea that once characterised these properties will be useful in understanding the genetics of complex traits in this cohort. This report hinges on information from a reconstructed multigenerational pedigree from Norfolk Island and available SNP genotype data for over 500 individuals comprising the pedigree. Pedigree and marker derived inbreeding coefficients were calculated using the IBDLD software, as well as runs of marker derived locus-specific homozygosity. Homozygosity by descent (HBD) was used to assess the locus-specific patterns of inbreeding. Statistical association testing was performed to explore relationships between cardiovascular disease related endophenotypes and the identified inbreeding and homozygosity patterns.

Results: Calculation of inbreeding based on the reconstructed pedigree structure revealed an average inbreeding coefficient of 0.011, a relationship that lies between second cousins and second cousins once removed. A combination of pedigree and SNP information was used to identify an overall prevalence of consanguinity in the genotyped core-pedigree individuals of 87%, with a mean marker-generated inbreeding coefficient of 0.011, which is consistent with the pedigree derived estimate. Numerous runs of HBD were identified across the genotyped core-pedigree individuals, with large regions of homozygosity observed on chr 6 (in the HLA region). One of the primary goals of identifying these multiple levels of genomic structure was to be able to make use of them in the search for potential markers of complex disease. Statistically significant correlations were observed between several CVD risk traits and inbreeding. The strongest correlation was observed between an aggregate risk score for CVD and marker derived inbreeding coefficient (r=0.389, P<2.4x10-11). Moreover, locus-specific HBD associations were observed for several CVD risk traits and SNPs localising to chr 5, 6 and 11 (P<0.05). 

Conclusions: This analysis has further characterised the unique population structure forged in the Norfolk population over the last 200 years. Estimates for inbreeding are ~10% and a new set of measurable indices that represent both runs of homozygosity have been added. Exploratory analyses comparing homozygosity indices with CVD risk traits suggest that individuals with close genetic ties to the original founders may have a potential predisposition to CVD. These findings will be important for future genetic analysis of complex disorders and indicate that homozygosity mapping may shed light on CVD risk in this cohort.