Application of weighted gene co-expression network analysis to identify novel key genes in diabetic nephropathy

Diabetic nephropathy (DN) is among the leading causes of end-stage renal disease worldwide. DN pathogenesis remains largely unknown. Weighted gene co-expression network analysis is a powerful bioinformatic tool for identifying key genes in diseases.

The datasets GSE30122, GSE104948, GSE37463 and GSE47185 containing 23 DN and 23 normal glomeruli samples were obtained from the National Center for Biotechnology Information Gene Expression Omnibus database. After data pre-processing, weighted gene co-expression network analysis was carried out to cluster significant modules. Then, Gene Set Enrichment Analysis-based Gene Ontology analysis and visualization of network were carried out to screen the key genes in the most significant modules. The connectivity map analysis was carried out to find the significant chemical compounds. Finally, some key genes were validated in in vivo and in vitro experiments.

A total of 454 upregulated and 392 downregulated genes were identified. A total of 16 modules were clustered, and the most significant modules (green, red and yellow modules) were determined. The green module was associated with extracellular matrix organization, the red module was associated with immunity reaction and the yellow module was associated with kidney development. We found several key genes in these three modules separately, and part of them were validated in vivo and in vitro successfully. We found the top 15 chemical compounds that could perturb the overall expression of key genes in DN.

Weighted gene co-expression network analysis was applied to DN expression profiling in combination with connectivity map analysis. Several novel key genes and chemical compounds were screened out, providing new molecular targets for DN.