Congrats to Jiwoo, who is first author of two independent papers published in Scientific Reports this month.
#1. The first paper describes disease biology learned from non-pleiotrophic variants between rheumatoid arthritis and systemic lupus erythematosus:
Title: Genetic variants differentially associated with rheumatoid arthritis and systemic lupus erythematosus reveal the disease-specific biology.
Abstract Two rheumatic autoimmune diseases, rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE), have distinct clinical features despite their genetic similarities. We hypothesized that disease-specific variants exclusively associated with only one disease could contribute to disease-specific phenotypes. We calculated the strength of disease specificity for each variant in each disease against the other disease using summary association statistics reported in the largest genome-wide association studies of RA and SLE. Most of highly disease-specific associations were explained by non-coding variants that were significantly enriched within regulatory regions (enhancers or H3K4me3 histone modification marks) in specific cell or organ types. (e.g., In RA, regulatory T primary cells, CD4+ memory T primary cells, thymus and lung; In SLE, CD19+ B primary cells, mobilized CD34+ primary cells, regulatory T primary cells and monocytes). Consistently, genes in the disease-specific loci were significantly involved in T cell- and B cell-related gene sets in RA and SLE. In summary, this study identified disease-specific variants between RA and SLE, and provided statistical evidence for disease-specific cell types, organ and gene sets that may drive the disease-specific phenotypes.
#2. The second papers describes our new computational program to infer disease signals at HLA variants by using SNP-disease association signals:
Title: Understanding HLA associations from SNP summary association statistics.
Abstract Strong genetic associations in the region containing human leukocyte antigen (HLA) genes have been well-documented in various human immune disorders. Imputation methods to infer HLA variants from single nucleotide polymorphism (SNP) genotypes are currently used to understand HLA associations with a trait of interest. However, it is challenging for some researchers to obtain individual-level SNP genotype data or reference haplotype data. In this study, we developed and evaluated a new method, DISH (direct imputing summary association statistics of HLA variants), for imputing summary association statistics of HLA variants from SNP summary association statistics based on linkage disequilibria in Asian and European populations. Disease association Z scores in DISH were highly correlated with those from imputed HLA genotypes in null model datasets (r = 0.934 in Asians; r = 0.960 in Europeans). We applied DISH to two previous GWAS datasets in Asian systemic lupus erythematosus and European rheumatoid arthritis populations. There was a high correlation between Z scores in the DISH and HLA genotype imputations, showing the same disease-susceptible and protective alleles. This study illustrated the usefulness of the DISH method in understanding and identifying disease-associated HLA variants in human diseases while maintaining individual-level data security.