Endocrine Abstracts

Type 2 diabetes (T2D) is a multi factorial disease with increased prevalence globally. The risk of developing T2D is determined by the interaction of genetic and environmental factors. Over the decades of research, hundreds of T2D genetic susceptibility markers (mostly single nucleotide polymorphisms or SNP) have been discovered. Yet generally their contribution to increased T2D risk is small and have negligible value in clinical practice to aid T2D treatment, categorize patients or guide prognosis. To overcome the weakness of individual genetic markers, polygenic risk score (PRS) method has been developed to account for genetic background of an individual and presence of multiple T2D risk markers. T2D associated SNPs vary across populations and countries which translate into even higher heterogeneity of a PRS model. Therefore, it is important to develop PRS model for each country and population. This study performed a genome wide association study in Latvian population using 822 cases and 802 controls and developed a PRS model to capture genetic risk of T2D in Latvian population. The results indicate that seven SNPs have a statistically significant association with T2D in Latvian population after Benjamini-Hochberg false discovery rate correction and considering population stratification and other covariates. The overall correlation of effects of SNPs with recent report in meta-analysis of European populations (Mahajan 2018) is high, reaching r = 0.39 when considering SNPs up to P = 1 × 10^–2. PRS model included 41 independent SNPs and contributes to 12% variability of T2D in the target group and 11% in the validation group with P = 1 × 10^–7 and AUC 0.69. Upper quartile individuals have T2D odds ratio 2.8 (CI 95% = 1.3–5.7) while lower quartile has odds ratio 0.3 (CI 95% = 0.3–0.7) when compared to the second quartile. Heritability of T2D has been determined at 26% in populations of European descent, therefore, capture of 11% variability translates to 42% of inherited variation explained in Latvian population by current PRS model. Many of the established T2D genetic risk factors were statistically significant (such as TCF7L2 gene) but failed to reach genome wide significance after correction for multiple testing, indicating that study is undepowered and larger sample size is required to improve detection of individual SNP effects and to improve PRS model. This T2D study has produced a strong PRS model explaining 11% of the diabetes variability in Latvian population and could be used in analysis with other T2D risk factors.