Meta-analysis of genome-wide association study identifies FBN2 as a novel locus associated with systemic lupus erythematosus in Thai population

Differences in the expression of variants across ethnic groups in the systemic lupus erythematosus (SLE) patients have been well documented. However, the genetic architecture in the Thai population has not been thoroughly examined. In this study, we carried out genome-wide association study (GWAS) in the Thai population.

We demonstrated the genetic architecture of SLE in the Thai population and identified a novel locus associated with SLE. Also, our study suggested a potential use of the PRS model from the Chinese population to estimate the disease risk for individuals of Thai ancestry.

Two GWAS cohorts were independently collected and genotyped: discovery dataset (487 SLE cases and 1606 healthy controls) and replication dataset (405 SLE cases and 1590 unrelated disease controls). Data were imputed to the density of the 1000 Genomes Project Phase 3. Association studies were performed based on different genetic models, and pathway enrichment analysis was further examined. In addition, the performance of disease risk estimation for individuals in Thai GWAS was assessed based on the polygenic risk score (PRS) model trained by other Asian populations.

Previous findings on SLE susceptible alleles were well replicated in the two GWAS. The SNPs on HLA class II (rs9270970, A>G, OR = 1.82, p value = 3.61E-26), STAT4 (rs7582694, C>G, OR = 1.57, p value = 8.21E-16), GTF2I (rs73366469, A>G, OR = 1.73, p value = 2.42E-11), and FAM167A-BLK allele (rs13277113, A>G, OR = 0.68, p value = 1.58E-09) were significantly associated with SLE in Thai population. Meta-analysis of the two GWAS identified a novel locus at the FBN2 that was specifically associated with SLE in the Thai population (rs74989671, A>G, OR = 1.54, p value = 1.61E-08). Functional analysis showed that rs74989671 resided in a peak of H3K36me3 derived from CD14+ monocytes and H3K4me1 from T lymphocytes. In addition, we showed that the PRS model trained from the Chinese population could be applied in individuals of Thai ancestry, with the area under the receiver-operator curve (AUC) achieving 0.76 for this predictor. Conclusions: We demonstrated the genetic architecture of SLE in the Thai population and identified a novel locus associated with SLE. Also, our study suggested a potential use of the PRS model from the Chinese population to estimate the disease risk for individuals of Thai ancestry.