Three common recessive variations explain more than 20% of all cases of type 2 diabetes in Greenland. A. Albrechtsen1, I. Moltke2, M. E. Jørgensen3, P. Bjerregaard4, E. V. R. Appel5, R. Nielsen6, O. Pedersen5, N. Grarup5, T. Hansen5 1) The Bioinformatics Centre, University, Copenhagen, Copenhagen, Denmark; 2) Department of Human Genetics, University of Chicago, Chicago, IL, USA; 3) Steno Diabetes Center, Gentofte, Denmark; 4) National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark; 5) The Novo Nordisk Foundation Center for Basic Metabolic Research, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark; 6) Department of Integrative Biology, University of California, Berkeley, CA, USA.

   Background and aims: We recently identified a common genetic variant in TBC1D4, which strongly influence plasma glucose levels. This variant also explained more than 10% of all cases of type 2 diabetes in Greenland and showed a recessive inheritance pattern. Motivated by this finding we analyzed Illumina metaboChip data for association to type 2 diabetes in Greenland applying a recessive model. Methods and Results: We performed GWAS and subsequent meta-analysis of two Greenlandic cohorts consisting in total of 3000 individuals that were either normal glucose tolerant or had type 2 diabetes. Association testing was done using a linear mixed model to control for admixture and relatedness. Besides the TBC1D4 locus, we identified two additional variations that under a recessive model are associated to type 2 diabetes (P 5×10-8 ). They are located on 22q12.3 and 5q1.2 far from known metabolic loci. Both variations are so common in the Inuit ancestry (MAF: 19.2% & 33.4%) that they have a strong impact under a recessive model. In contrast, they are not as common in Europeans (MAF: 7.3% & 3.3%) and therefore do not show a large recessive impact on the population. The type 2 diabetes prevalence in the Greenlandic cohorts is 10.1%, however homozygous carriers of the two identified variants have 30.1% and 24.0% risk of type 2 diabetes, respectively. Homozygous carriers of the previously identified TBC1D4 variant have a risk of 54.3%. A little more than 10% of the Greenlandic population is homozygous carriers of at least one of the 3 variants, which increases their probability of having type 2 diabetes to 25.8% compared to a risk of 8.5%. This corresponds to a population attributable risk of 24.1%. Conclusion: We have identified two additional recessive variants, each with a large impact on type 2 diabetes risk in Greenland that would not have been found applying an additive model. Together, the two identified variants and the previously identified TBC1D4 variant explains more than 20% of all cases of type 2 diabetes in the Greenlandic population. These findings illustrate that it can be valuable to also apply a recessive model when performing GWAS, especially in historically small and isolated founder populations, like the Greenlandic.

You may contact the first author (during and after the meeting) at