Mapping Multiple Sclerosis Risk Across Diverse Ancestries

Multiple sclerosis (MS) is a chronic inflammatory and neurodegenerative disorder whose susceptibility is shaped by a highly polygenic genetic architecture. Large genome-wide association studies (GWAS) have identified more than 200 susceptibility loci, with the strongest effects concentrated in the major histocompatibility complex (MHC), particularly among human leukocyte antigen (HLA) genes. However, most discoveries have emerged from populations of European ancestry, creating substantial uncertainty about whether the same variants, effect sizes and biological mechanisms operate in other populations. The study Genetic Determinants of Multiple Sclerosis Susceptibility in Diverse Ancestral Backgrounds addresses this imbalance by examining MS genetics in UK residents of South Asian and African genetic ancestry. The investigation is particularly important because limited ancestral diversity can reduce the accuracy and clinical portability of genetic risk scores, impede the fine-mapping of causal variants and reinforce inequalities in genomic medicine. The authors therefore sought not only to identify susceptibility associations in underrepresented populations, but also to quantify the extent to which the genetic architecture established in European cohorts is shared across ancestral groups.

Constructing an Ancestrally Diverse Genetic Cohort
The research was conducted through the ADAMS project—A Genetic Association Study in Diverse Ancestries of Multiple Sclerosis—which recruited people with self-reported or clinically diagnosed MS through clinical centres, an online platform, primary care services and the UK MS Register. Participants supplied saliva samples for DNA extraction and completed standardized questionnaires covering demographics, disease history, treatment exposure, established environmental risk factors and disability-related outcomes. Genotyping was performed with the Illumina Global Screening Array, followed by quality control and genotype imputation. The case cohort was then integrated with ancestrally comparable controls from UK Biobank. Genetic ancestry was inferred through principal-component-based classification using reference populations from the Human Genome Diversity Project and the 1000 Genomes Project, rather than being determined solely from self-reported ethnicity. After ancestry-specific filtering and quality control, the principal GWAS datasets contained 175 MS cases and 6,744 controls of South Asian ancestry and 113 cases and 5,177 controls of African ancestry. Association testing used mixed logistic models adjusted for sex and ten genetic principal components, thereby reducing confounding from residual population structure.

The MHC Remains the Dominant Susceptibility Region
The ancestry-specific GWAS confirmed that the MHC on chromosome 6 remains the most prominent genetic region associated with MS susceptibility across populations. In the South Asian cohort, the leading signal was located in the class II HLA region near HLA-DRB1. The lead variant, chr6:32600515:G:A, was more frequent among cases than controls and was associated with an estimated odds ratio of 1.84. In the African-ancestry analysis, the strongest signal arose in the class I region near HLA-A; chr6:29919337:A:G produced an estimated odds ratio of 2.24. Several additional loci outside the MHC reached nominal or suggestive thresholds, but none met the conventional genome-wide significance criterion of P < 5×10−8. The authors appropriately interpret these non-MHC findings as likely products of statistical noise, limited power or residual stratification rather than confirmed ancestry-specific susceptibility loci. Importantly, genome-wide test statistics showed no substantial inflation, suggesting that the analytical procedures controlled population structure effectively. Thus, the study does not establish new genome-wide significant loci, but it provides independent evidence that MHC-mediated immune regulation is central to MS susceptibility in both South Asian and African ancestral backgrounds.

Shared Architecture with Ancestry-Dependent Effect Estimates
To determine whether previously identified European-ancestry risk variants operate similarly in other populations, the researchers examined 164 independent susceptibility signals derived from a large European GWAS. Among South Asian participants, 104 of 154 testable variants showed the same direction of effect as in the European analysis, producing statistically significant directional concordance and a moderate Spearman correlation of approximately 0.31. In the African-ancestry cohort, 80 of 152 variants showed concordant effects, but the overall enrichment and correlation were weaker and did not reach statistical significance. Excluding the extended MHC region did not eliminate the pattern, indicating that cross-ancestry similarity was not driven exclusively by HLA variation. Nevertheless, the correlations remained below 0.5, demonstrating that shared biological pathways do not necessarily imply identical marginal effect estimates. Differences in linkage disequilibrium, allele frequency, local haplotype structure and statistical precision can all alter the apparent association between a measured marker and the underlying causal variant. When the analysis was restricted to variants with more informative association statistics, evidence of directional consistency became stronger in both ancestral groups. The findings therefore support a model in which much of the biological architecture of MS is shared, while the observable genomic signatures of that architecture vary among populations.

HLA Fine-Mapping Reveals Convergence and Possible Population-Specific Signals
Classical HLA alleles were imputed at six loci and evaluated through two independent imputation frameworks, enabling a more detailed examination of the MHC associations. In South Asian participants, the analysis identified risk-increasing associations involving HLA-DPB1*10:01, HLA-B*37:01, HLA-A*26:01, HLA-DRB1*15:01, HLA-A*23:01 and HLA-DRB1*04:01, together with potentially protective associations for HLA-DRB1*13:01 and HLA-DQB1*06:03. In the African-ancestry cohort, HLA-A*66:01 was the only allele to pass the study’s 10% false-discovery-rate threshold, although nominal associations were also observed for alleles including HLA-DRB1*15:01. Several associations corresponded to known European-ancestry signals, reinforcing the existence of shared immunogenetic mechanisms. Others, particularly HLA-DPB1*10:01, HLA-A*26:01, HLA-A*23:01 and HLA-A*66:01, may represent population-enriched effects but require independent replication. The study also illustrates why allele frequency must be distinguished from individual-level effect: HLA-DRB1*15:01 increased risk across ancestries, yet its lower frequency in South Asian and African populations resulted in estimated population-attributable fractions of approximately 9.8% and 4.5%, respectively, compared with a substantially larger estimate in European-ancestry populations.

Reduced Portability of European-Derived Polygenic Risk Scores
The authors further assessed cross-ancestry transferability by constructing polygenic risk scores from European-ancestry GWAS summary statistics. In both study populations, MS cases generally carried higher scores than controls, and the proportion of cases tended to increase across progressively higher risk-score quartiles. However, the predictive performance was markedly reduced relative to previous European-ancestry analyses. The best-performing score explained approximately 1.6% of MS liability in South Asian participants, compared with about 4.3% in an earlier European-ancestry UK Biobank analysis. In the African-ancestry cohort, the score explained only 0.5% of liability, with an empirical P-value of 0.08, indicating weak evidence under conventional significance criteria. This decline does not imply that fundamentally different diseases occur in different populations. Rather, it reflects the dependence of polygenic scores on ancestry-specific allele frequencies, linkage disequilibrium patterns, imputation quality and effect-size estimation. The results demonstrate that applying European-derived scores without recalibration may generate systematically less informative predictions for underrepresented groups. At present, such scores should therefore be interpreted as population-level research instruments rather than ancestry-neutral clinical tests. The observed monotonic relationship between score quartiles and MS prevalence nevertheless indicates that European-derived scores capture a limited component of shared susceptibility, particularly in the South Asian cohort.

Limitations, Scientific Significance and Future Priorities
The study’s conclusions must be interpreted in light of several important limitations. The most consequential is sample size: although this represents a major UK effort in diverse-ancestry MS genetics, the numbers are small by modern GWAS standards and provide insufficient power to establish modest or rare variant associations. Cases and controls were also obtained largely from different cohorts, creating potential batch effects related to DNA source, genotyping platform, imputation and demographic composition. The investigators employed conservative variant filtering, joint re-imputation, ancestry-specific quality control and mixed-model association testing, but no highlighted result achieved genome-wide significance. HLA alleles were inferred computationally rather than determined through sequencing or PCR-based typing, and some participants were excluded because of strict ancestry definitions. Moreover, the manuscript is a preprint and had not been certified by peer review at the time of posting. Despite these constraints, the study provides persuasive evidence that the MHC and many established susceptibility pathways are relevant across ancestral backgrounds, while also showing that allele frequencies, association patterns and predictive performance remain population dependent. The authors argue that future studies involving tens of thousands of ancestrally diverse participants could improve causal fine-mapping, identify variants uncommon in European populations, enhance equitable risk prediction and reveal therapeutically actionable mechanisms.

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:
Jacobs, B. M., Schalk, L., Tregaskis-Daniels, E., Scalfari, A., Nandoskar, A., Dunne, A., ... & Dobson, R. (2026). Genetic determinants of multiple sclerosis susceptibility in people from diverse ancestral backgrounds. Neurology, 106(7), e214708.