Abstract Text: Multiple sclerosis (MS), an autoimmune disease of the nervous system affecting 2.8 million individuals worldwide, has well-established genetic risk factors. However, mechanisms mediating these genetic effects and their interactions with MS environmental risk factors remain unclear. To better understand immune-related diseases like MS, we developed a novel approach to flexibly identify fine-grained immune cell states that increase in abundance with the allelic dose of a genetic variant. When applied in a genome-wide survey, our method identifies a spectrum of immune states under genetic influence, building on previous studies using flow cytometry. However, our approach uses single-cell data without pre-specifying candidate cell types in order to characterize genetically-associated cell states with more flexibility and statistical power. In data from 969 Australians of European ancestry (the OneK1K cohort), our method reveals >150 independent genome-wide significant loci associated with changes in T, B, NK and Myeloid cell populations, including individual variants associated with simultaneous shifts across multiple cell types. One of these loci, in the major histocompatibility complex region, contained the strongest associations genome-wide. Fine-mapping within this region reveals that most of this MHC-region signal is explained by HLA-DRB1*15:01, the major MS genetic risk allele. We characterize shifts in B (P=4e-182), T (P=1e-14) and dendritic cell (P=2e-16) population abundances associated with increasing allelic dose of HLA-DRB1*15:01. These shifts appear in individuals without MS disease, and may illuminate how HLA-DRB1*15:01 interacts with pathogenic exposures like EBV to set the immunological stage for MS.
