Integrative Analysis of Transcriptome-Wide Association Study and Gene-Based Association Analysis Identifies Candidate Genes Associated With Juvenile Idiopathic Arthritis

Background: The current genome-wide association study (GWAS) of Juvenile idiopathic arthritis (JIA) suffers from the low power due to limited sample size, as well as suffers from the interpretation issue due to most GWAS signals located in non-coding regions. Gene-level analysis is able to aggregate many SNPs with small effects to improve the association power, and further benefit the better understanding of genetic determinants underlying JIA. Methods: Using the largest GWAS summary statistics of JIA to date (3305 cases and 9196 controls), we performed transcriptome-wide association studies (TWAS) using FUSION, parallelized with gene-based association analysis using the eMAGMA method, to identify the potential tissue-specific genes related with JIA. We then explore the novel JIA-associated genes through overlapping the genes significantly detected from these two typical gene-level analyses to avoid the risk of false discoveries from using single method, followed by enrichment analysis to identify the significant gene ontology terms as well as the pathways. Results: A total of 33 unique genes had been significantly identified from both TWAS analysis and eMAGMA gene-based association analysis, of which 11 were previously reported, including TYK2 ( P FUSION = 5.12 × 10 -6 , P eMAGMA = 1.94 × 10 -7 for Whole Blood), IL(Interleukin)-6R ( P FUSION = 8.63 × 10 -7 , P eMAGMA = 2.74 × 10 -6 for Cells EBV-transformed lymphocytes) and Fas ( P FUSION = 5.21 × 10 -5 , P eMAGMA = 1.08 × 10 -6 for Muscle Skeletal). There are totally 22 newly reported genes indicating the power advantage of gene-level association analysis, of which some are more likely JIA-associated genes, including IL-27 ( P FUSION = 2.10 × 10 -7 , P eMAGMA = 3.93 × 10 -8 for Liver), LAT ( P FUSION = 1.53 × 10 -4 , P eMAGMA = 4.62 × 10 -7 for Artery Aorta) and MAGI3 ( P FUSION = 1.30 × 10 -5 , P eMAGMA = 1.73 × 10 -7 for Muscle Skeletal). Enrichment analysis of 33 common genes further implicated the significant roles of 10 GO terms as well as 4 KEGG pathways including Th17 cell differentiation ( P= 5.83×10 -6 ) and Rap1 signaling pathway ( P= 1.55×10 -3 ). Conclusions: Our findings provide novel insights into the genetic determinants of JIA, which could benefit the understanding of pathogenic mechanisms as well as potential therapeutic targets of JIA.