Background/Purpose:  Rheumatoid arthritis (RA) affects 0.5-1% of the population worldwide. The genetics of RA has been analyzed in large European and Asian studies, but much less is known about RA in African-Americans. Trans-ethnic analysis has found similarities and differences in RA genetics, making both validation and novel gene association important goals in African-Americans (AA). We present genome-wide association (GWA) data on African-Americans with RA alongside analysis of whole genome sequencing (WGS) data on a subset of the population. These data were analyzed in the context of publically available datasets on RA.

Methods:  535 AA cases (from the CLEAR Registry and the VARA Registry) and 1,506 AA controls (from CLEAR and SLEGEN) were genotyped using Illumina Omni 1M and 1S platforms. Inclusion thresholds were: sample call rate > 98.5%, SNP call rate > 98%, HWE p-value > 1×10^-7, MAF >0.05.  The top principal component was included as a covariate. Imputation was carried out using IMPUTE2 with 1000 Genomes. The number of SNPs after merging was 1,790,756, and the total number of pruned SNPs after imputation was 8,380,261. After this, genomic inflation was 1.031. Associated and suggestive loci were defined as 500kB on either side of lead SNPs with a p-value of <10^-8, or < 10^-5, respectively. We gathered available GWAS, transcriptomic, methylomic data, and other data in RA for these loci. Next, we integrated this with analysis of copy number, structural, and rare variation (CNV, SV and RV) from WGS of 62 AA RA patients who were also genotyped on the Illumina platform and, 80 Yoruban (YRI), and 96 Northern European (CEU) genomes sequenced by CGI. We annotated rare variants with Combined Annotated Dependent Depletion (CADD) scores as well as according to 27 other ontologies to aid interpretation.

Results:  One associated and 44 suggestive loci were identified among African-Americans with RA through GWA analysis. Two of these loci were previously associated with RA and another ten have been identified in other GWAS on autoimmune diseases or immunologic conditions. Multiple CNV were identified in the regions circumscribed by GWAS, some of which were several times more common in genomes of cases than controls. We found neither large deletions nor insertions unique to the RA population in the loci. Gene Burden Testing (GBT) reaffirms prior associations (e.g. SYNGR1 displays 2.71 times more RV per genome than a weighted average of the CEU and YRI genomes). GBT appears particularly effective at enabling identification of associated antisenseRNAs and microRNAs. Most importantly, integrated analysis of our data displays internal consistency; for instance SEMA6D is in a locus defined by GWAS, is differentially methylated in RA, and is identified by GBT using WGS.

Conclusion: This co-analysis of GWAS, WGS, and publically available data allows refinement of gene and variant-level associations. This study re-confirms the importance of CNV in RA, but in loci not yet reported to have such. Finally, our results show that GBT enables confirmation and refinement of associations as well as the identification of RV likely to be deleterious in the context of RA.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/genetic-influences-on-rheumatoid-arthritis-in-african-americans/