Aim Multiple sclerosis (MS) is a neurodegenerative disorder for which the primary risk allele, HLA-DRB1*15:01, was identified a half century ago. Nevertheless, multiple additional independent association signals have been mapped via single nucleotide polymorphism (SNP) studies to the extended major histocompatibility (MHC) region. However, the high levels of polymorphism, structural variation, and linkage disequilibrium characteristic of the MHC have confounded efforts to identify the true risk loci associated with these SNP signals. Further, few studies have investigated MS risk at the MHC in African American populations. We aim to more clearly resolve the architecture of MS risk in the MHC, in both European-American and African-American populations. Methods We applied targeted next-generation sequencing (NGS) to the entire extended 5Mb MHC to comprehensively detail sequence variation across the region in two cohorts of MS cases and controls of European American (n=382) and African American (n=170) ancestry. Using our novel pipeline, MHConstructor, we generated reference-free, de novo MHC sequence assemblies and performed association testing for disease across the entire region. Results We find five strong MS risk association signal peaks that occur in the same genomic locus in both the European American and African American cohorts, allowing definitive assignment of risk to these loci: SCAND3 upstream, an intron of LINC02571, the intergenic region between HLA-F and HLA-G, a TSBP1 intron and SMIM40 upstream region. TSBP1 and HLA-G have been previously implcated in MS risk, and the LINC02571, TSBP1 and HLA-G upstream regions overlap with loci previously identified to be differentially methylated in MS. Conclusion By examining variation across the extended MHC in this trans ancestral analysis, we provide the most high-resolution picture of MS risk variation within the MHC to-date, a critical step towards determining the full genetic underpinning of MS pathogenesis.
