Kazakh genetic risk scores transfer unevenly, study finds

Polygenic scores built to predict coronary artery disease and hypertension do not transfer equally well into an underrepresented population, according to a new analysis of a Kazakh cohort with ischemic heart disease and co-occurring arterial hypertension. The study, published in BMC Research Notes, tested four previously published polygenic scores in 470 ethnic Kazakh individuals and found that only one retained statistical significance after correction for multiple testing. The work examined how published polygenic scores behave in a population underrepresented in their development datasets.

The research team evaluated the scores in a case-control design involving 240 individuals with ischemic heart disease and arterial hypertension and 230 controls. Three of the four scores showed an association with disease status after adjustment for age, sex and ten genetic principal components. However, when the researchers corrected for multiple testing, only one score, PGS005144, remained significant, with an odds ratio of 1.33 per standard deviation increase and a 95 percent confidence interval of 1.10 to 1.62. Another score, PGS003018, showed no association at all, with an odds ratio of 1.01 and a confidence interval spanning 0.84 to 1.23.

The Kazakh study was designed to explore exactly these technical and population-genetic factors.

To understand why the scores performed differently, the researchers examined the 100 highest-weighted variants in each score. For PGS005144, 90 of these variants were retained after imputation and quality control, representing 91.4 percent of the summed absolute effect weight. For PGS003018, only 73 variants were retained, representing 75.4 percent of the summed absolute effect weight. The clearest difference between the two scores was poorer recovery of highly weighted PGS003018 variants after imputation and quality control.

The researchers also investigated whether ancestry-specific genetic architecture could explain the differences. They performed allele-frequency comparisons and local linkage disequilibrium analyses. These analyses did not support a simple East Asian-versus-European explanation for the observed differences in score transferability.

Internal validation within the cohort did not show clear improvement in disease discrimination when the polygenic score was added to existing clinical or demographic predictors.

The study evaluated four published polygenic scores in a Kazakh case-control cohort, a population underrepresented in the scores' development datasets.

Subject of Research: Transferability of published coronary artery disease and hypertension polygenic scores in a Kazakh cohort

Article Title: Transferability of published coronary artery disease and hypertension polygenic scores in a Kazakh cohort with ischemic heart disease and co-occurring arterial hypertension

Article References: Skvortsova, L., Yergali, K., Zhaxylykova, A., Begmanova, M., & Mansharipova, A. (2026). Transferability of published coronary artery disease and hypertension polygenic scores in a Kazakh cohort with ischemic heart disease and co-occurring arterial hypertension. BMC Research Notes. https://doi.org/10.1186/s13104-026-08091-z

Image Credits: AI Generated

DOI: 10.1186/s13104-026-08091-z

Keywords: polygenic scores, coronary artery disease, hypertension, Kazakh population, ischemic heart disease, genetic risk prediction, population genetics, imputation quality, linkage disequilibrium, allele frequency, transferability, BMC Research Notes