Data Integration for Big Data analytics to identify the gaps in Rheumatoid Arthritis Genomics in a Post-GWAS era

Genome-wide association studies (GWASs) enable simultaneous detection of several genetic variations. The variants that do not match GW ASs' significance threshold can be missed. We aimed at the identification of novel candidate single nucleotide polymorphisms (SNPs) for rheumatoid arthritis (RA) using data integration based bioinformatics for analysis of publically available big data that employ computational processing and machine learning. 340 GW AS significant RA SNPs were searched for the proxy SNPs using SNipA proxy search tool that were then scored and annotated. Of 9194 proxy SNPs, 492 SNPs returned significant RegulomeDB scores and were linked to 94 expression quantitative trait loci (eQTLs). Protein interaction analysis identified 13 ‘hub proteins’ expressed by these eQTLs linked to 54 SNPs. Of these, nine were already GW AS significant RA SNPs. Remaining 45 novel SNPs, mapped to 11 genomic loci, are novel candidate SNPs for RA. The proposed pipeline can be integrated in an automated machine learning algorithm to automate the identification of novel SNPs for other disease with underlying complex genomics.