Integrated analysis of different microarray studies to identify candidate genes in type 1 diabetes

Background Type 1 diabetes (T1D), an autoimmune disease, occurs most commonly in children. Identifying altered gene expression in peripheral blood mononuclear cells (PBMCs) of T1D may lead to new strategies for preserving or improving β-ell function in patients with T1D. Methods The Gene Expression Omnibus database was searched for microarray studies in PBMCs of T1D. Subsequently, gene expression datasets from multiple microarray studies were integrated to obtain differentially expressed genes (DEGs) between T1D and normal controls (NC). Gene function analysis was performed to determine the functions of the DEGs identified. Results Four microarray studies were available for analysis, including 199 T1D samples and 74 NC samples. Analysis revealed 695 genes that were significantly differentially expressed in PBMCs from T1D compared with NC samples, with 450 upregulated and 245 downregulated. Signal transduction (gene ontology [GO]: 0007165; false discovery rate [FDR] = 1.54 × 10–7) and protein binding (GO: 0005515; FDR = 2.93 × 10–24) were significantly enriched for the GO categories of biological processes and molecular functions, respectively. The most significant pathway in the Kyoto Encyclopedia of Genes and Genomes analysis was arachidonic acid metabolism (FDR = 1.44 × 10–3). Protein–protein interaction network analysis showed that the significant hub proteins contained immature colon carcinoma transcript 1 (ICT1; degree = 214; clustering coefficient [C] = 4.39 × 10–5), zinc finger and BTB domain containing 16 (ZBTB16; degree = 112; C = 8.04 × 10–4), and SERTA domain containing 1 (SERTAD1; degree = 38; C = 0.0014). Conclusions This integrated analysis will help develop improved therapies and interventions for T1D by identifying novel drug targets.