Assessing the Limits of European-Derived Polygenic Risk Scores for Multiple Sclerosis Prediction in South Asian Populations
The article, “Polygenic risk score prediction of multiple sclerosis in individuals of South Asian ancestry,” examines whether genetic risk models developed mainly from European-ancestry genome-wide association studies can accurately predict multiple sclerosis risk in people of South Asian ancestry. This question is highly important because polygenic risk scores are increasingly discussed as tools for disease prediction, preventive trial design, and early risk stratification, yet their performance may vary substantially across populations.
Multiple Sclerosis as a Polygenic Disease
Multiple sclerosis is a complex immune-mediated neurological disorder influenced by many genetic variants, each usually contributing a small amount to total disease risk. The strongest genetic contribution comes from the major histocompatibility complex region, especially human leukocyte antigen variation, but more than 200 additional non-MHC susceptibility loci have also been identified. The study therefore focuses on whether aggregating these risk variants into a polygenic risk score can meaningfully distinguish individuals with multiple sclerosis from controls across different ancestry groups.
Study Design and Cohorts
The investigators used two major longitudinal genetic resources: Genes & Health, which includes British–Bangladeshi and British–Pakistani individuals, and UK Biobank, which is predominantly composed of individuals of European ancestry. After quality control, the Genes & Health analysis included 40,532 South Asian-ancestry participants, of whom 42 had multiple sclerosis, while the European-ancestry UK Biobank comparison included 2,091 cases and 374,866 controls. This design allowed the authors to compare the portability of European-derived multiple sclerosis polygenic risk scores between ancestry groups.
Polygenic Risk Score Construction
The authors calculated multiple polygenic risk scores using summary statistics from the largest multiple sclerosis genome-wide association study available at the time. They applied a clumping-and-thresholding strategy, varying linkage disequilibrium thresholds and association P-value thresholds to identify optimal scores. Importantly, they generated scores both including and excluding the MHC region, because this region has a dominant role in multiple sclerosis genetics and may behave differently across ancestries due to variation in linkage disequilibrium structure.
Findings in the South Asian-Ancestry Cohort
In the South Asian-ancestry Genes & Health cohort, the European-derived polygenic risk scores showed only modest predictive value. The optimal score including the MHC explained approximately 1.1% of liability to multiple sclerosis, while the score excluding the MHC explained approximately 1.5%. The visual results on page 5, including density plots, receiver operating characteristic curves, and calibration plots, show that the scores had some discriminatory capacity but that much of the apparent classification performance was also explained by covariates such as age, sex, and genetic principal components.
Comparison with European-Ancestry Participants
The same approach performed better in European-ancestry UK Biobank participants. In the full European-ancestry cohort, the MHC-containing score explained approximately 4.4% of multiple sclerosis liability, while the non-MHC score explained approximately 2.3%. Even after subsampling UK Biobank to match the smaller number of South Asian multiple sclerosis cases, the MHC-containing score still explained more liability in European-ancestry participants than in the South Asian-ancestry cohort. This supports the authors’ central conclusion that European-derived multiple sclerosis polygenic risk scores have reduced transferability to South Asian populations.
Interpretation and Broader Implications
The study highlights a major challenge for precision medicine: genetic prediction tools trained predominantly in European populations may underperform in other ancestral groups and thereby risk reinforcing existing health inequalities. The likely explanation is not that multiple sclerosis has completely different genetic causes across populations, but rather that allele frequencies and linkage disequilibrium patterns differ, reducing the accuracy with which European-discovered variants tag causal variation in South Asian genomes. The authors therefore argue that larger and more diverse genome-wide association studies are essential if polygenic risk scores are to become clinically or scientifically useful across global populations.
Disclaimer: This blog post is based on the provided research article and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.
References:
Breedon, J. R., Marshall, C. R., Giovannoni, G., van Heel, D. A., Dobson, R., & Jacobs, B. M. (2023). Polygenic risk score prediction of multiple sclerosis in individuals of South Asian ancestry. Brain Communications, 5(2), fcad041.
