Multiple Sclerosis Genetics Beyond Europe: Evidence for Shared HLA Risk and the Limits of Polygenic Transfer
Multiple Sclerosis (MS) has a substantial heritable component, and large genome-wide association studies (GWAS) in predominantly European-ancestry cohorts have already mapped hundreds of associated loci—most notably strong signals in the Major Histocompatibility Complex (MHC) region and many smaller effects elsewhere.
Yet the global relevance of these discoveries depends on whether the same risk architecture applies across ancestries, because unequal representation can translate directly into inequitable genetic prediction and an incomplete biological understanding of disease. This preprint addresses that gap by focusing on people of South Asian and African ancestry living in the UK, aiming to test overlap with European-ancestry findings and clarify how MS susceptibility signals behave in diverse genetic backgrounds.
Building the ADAMS cohort: a practical route to diversity in MS research
The authors assembled the ADAMS project (A Genetic Association study in Diverse Ancestries of Multiple Sclerosis), recruiting people with self-reported or clinically diagnosed MS through an online platform, clinical sites, primary care, and the UK MS Register. They paired this with structured phenotyping—capturing demographics, MS history, established risk factors (e.g., smoking history, glandular fever), and disability/quality-of-life measures such as EDSS-derived metrics and MSIS-29—so the genetic dataset is anchored to clinically meaningful characteristics rather than being “genotypes without context.”
From saliva to analyzable genomes: genotyping, imputation, and ancestry inference
DNA was collected using saliva kits and genotyped on the Illumina Global Screening Array, then imputed initially using the Haplotype Reference Consortium panel (via the Michigan Imputation Server) and later re-imputed using 1,000 Genomes in the final joint case-control datasets after additional batch-effect safeguards. A key methodological step was genetic ancestry inference: the team used PCA-based features and a random forest classifier trained on Human Genome Diversity Project and 1,000 Genomes reference data to identify ancestrally similar cases and controls, and to exclude ambiguous assignments to reduce population stratification risk.
Case-control design at scale: combining ADAMS with UK Biobank
Because assembling very large non-European MS case cohorts is difficult in any single study, the authors leveraged UK Biobank (UKB) as a source of ancestrally matched controls (and a small number of additional cases). After quality control and ancestry assignment, analyses focused on genetically inferred African-ancestry participants (112 ADAMS cases + 17 UKB cases; 7,701 UKB controls) and South Asian-ancestry participants (201 ADAMS cases + 13 UKB cases; 9,001 UKB controls). GWAS was performed within ancestry using mixed logistic models (REGENIE) adjusted for sex and principal components; age was not included in the primary models because it was highly collinear with case-control status given cohort differences between ADAMS and UKB.
Genome-wide signals: the MHC stands out, as expected
The most consistent finding was that the MHC region (chromosome 6) shows the clearest evidence of association with MS risk in both South Asian and African ancestry analyses—aligning with decades of MS genetics and reinforcing the idea that core immunogenetic mechanisms generalize across populations. Outside the MHC, the study reported a small number of suggestive loci at nominal thresholds (P < 1×10⁻⁴)—three in the South Asian analysis and eighteen in the African analysis—but importantly none of these reached genome-wide significance, and they did not cluster near known European-ancestry suggestive signals (outside the MHC). The responsible interpretation is that these are hypothesis-generating leads rather than confirmed new biology.
Zooming into the MHC: HLA signals largely align, with hints of population-specific effects
Within the MHC, the authors performed HLA imputation and sensitivity checks using two approaches (HIBAG and SNP2HLA), looking for consistency rather than relying on a single imputation model. The overarching pattern is concordance of effect direction for major European-ancestry HLA risk alleles—most notably HLA-DRB115:01—across South Asian and African ancestry cohorts, supporting shared disease pathways. At the same time, the strongest inferred effect in the African-ancestry cohort was HLA-A66:01, an allele enriched in African-ancestry participants in UK Biobank, raising the possibility (not proof) of ancestry-influenced signals that become visible when allele frequencies differ substantially between populations.
Clinical translation check: polygenic risk scores transfer, but with reduced performance
To quantify cross-ancestry overlap more directly, the team applied polygenic risk scores (PRS) derived from European-ancestry MS summary statistics to South Asian and African ancestry participants. The PRS separated cases from controls at a population level in both ancestries, but performance dropped compared with European benchmarks—consistent with known issues in PRS portability due to allele-frequency and linkage-disequilibrium differences. In their best models, the PRS explained ~1.6% of MS liability in the South Asian cohort and ~0.5% in the African cohort, versus a previously reported ~4.3% in European-ancestry UK Biobank analyses. The authors are explicit about limitations—small sample size, potential batch effects from combining distinct cohorts, reliance on imputation for HLA and on self-reported MS for some participants—and emphasize that none of the highlighted associations met genome-wide significance, framing this work as a foundational step toward larger multi-ancestry MS GWAS efforts that can enable robust discovery, fine-mapping, and fairer genetic prediction.
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. (2025). Genetic determinants of Multiple Sclerosis susceptibility in diverse ancestral backgrounds. medRxiv, 2025-01.
