Effect-Allele Frequencies at Cardiometabolic Risk Loci in an Uzbek Cohort from Tashkent: A Central Asian Reference and Priorities for Polygenic Score Validation.

Abstract

Background/objectives: Central Asian populations are almost entirely absent from genomic reference resources, yet Uzbekistan has one of the fastest rates of growth in age-standardised type 2 diabetes (T2D) incidence. Allele frequencies at cardiometabolic risk loci are unknown, limiting the interpretation of polygenic risk scores (PRS) derived elsewhere. We aimed to characterise effect-allele frequencies in an Uzbek cohort from Tashkent and quantify divergence from 1000 Genomes panels to guide future PRS evaluation.

Methods: We genotyped 171 adults from Tashkent (131 with T2D, 40 without) on a custom Agilent SureSelect panel of cardiometabolic candidate SNPs. After quality control, 1939 variants were analysed. Effect-allele frequencies were compared with 1000 Genomes Phase 3 superpopulations (EUR, SAS, EAS, AFR) using Hudson FST and Nei's genetic distance with bootstrap 95% CIs (1000 replicates). Case-control association was assessed descriptively using both unadjusted and age- and sex-adjusted additive logistic regression, with the adjusted model as primary.

Results: Reference frequencies were available for 1533 of 1939 variants (1363 at MAF ≥ 0.05). Frequencies clustered with South Asian (FST = 0.077, 95% CI 0.065-0.089) and European panels (FST = 0.086, 95% CI 0.072-0.099), which were indistinguishable (ΔFST = -0.010, 95% CI -0.022 to 0.003) and more divergent from East Asian and African panels. Divergence from the European and South Asian panels was unrelated to nominal association status (Mann-Whitney p = 0.99 and 0.57). In the primary age- and sex-adjusted analysis, no variant survived multiple-testing correction (minimum Bonferroni-adjusted and false-discovery-rate q = 0.28), consistent with the unadjusted sensitivity analysis.

Conclusions: This study provides the first systematic allele-frequency reference for cardiometabolic risk loci in an Uzbek cohort from Tashkent. The profile is closest to South Asian and European panels and identifies PRS from these ancestries as priority candidates for future local validation in Central Asian settings. The case-control analysis was underpowered and is reported descriptively.