Abstract

Polygenic risk scores (PRS) are increasingly used to estimate genetic susceptibility to complex diseases such as 
type 2 diabetes (T2D). However, most PRS models are derived from genome-wide association studies conducted 
primarily in European populations, raising concerns about their accuracy and clinical applicability in other 
ancestry groups. This paper reviews the transferability and predictive performance of PRS for T2D in South Asian 
populations, who experience a disproportionately high burden of the disease. It examines how differences in 
genetic architecture, allele frequencies, linkage disequilibrium structure, environmental exposures, and phenotypic 
heterogeneity influence PRS validity across populations. The review highlights evidence showing reduced 
discrimination, calibration issues, and potential bias when European-derived PRS are applied to South Asians. It 
further discusses methodological strategies to improve performance, including ancestry-specific genome studies, 
multi-ancestry modelling, recalibration approaches, and integration of PRS with clinical and lifestyle risk factors. 
Ethical and implementation considerations, such as equitable representation in genomic research, data governance, 
and responsible clinical deployment, are also addressed. The paper concludes that while PRS holds promise for 
enhancing early risk stratification and precision prevention of T2D in South Asians, substantial improvements in 
population diversity, validation, and clinical integration are required before routine use can be justified.