Validation of genetic risk scores for hypertension in the Central Russian population

Abstract

Aim. To validate and evaluate the accuracy of 4 genetic risk scores (GRSs) for hypertension (HTN), previously created on European samples, on a population sample of the Ivanovo Oblast.Material and methods. For genetic analysis, targeted next-generation sequencing was used on a sample of the Central Russia (n=1682) based on the biobank collection. Four GRSs associated with HTN, previously developed for the European population, were selected for validation. The coefficient of determination and the area under the ROC curve were used as quality metrics for regression models. Additional validation was carried out to include all nucleotide sequence variants, regardless of linkage disequilibrium level. A combined GRS was compiled based on coefficients from individual GRSs using the clumping + thresholding (C+T) method.Results. The study demonstrated that the predictive value of previously developed GRSs when used for Central Russian population is lower than in the original studies. The proportion of explained variance was 0,5-0,8%. The best predictive ability (proportion of explained variance — 2,5%) was demonstrated using previously developed GRSs (Evangelou E, et al., 2018), which includes the largest number of nucleotide sequence variants (n=852).Conclusion. GRSs for HTN, developed on European samples, is not recommended for Russian population without preliminary validation. To create original GRSs, combining statistical parameters (β-coefficients and p-value) from different GRS is not recommended.