Background/Aims: Podocyte damage is associated with proteinuria, glomerulosclerosis and decline of renal function. This study aimed to screen critical genes associated with podocyte injury in chronic kidney disease (CKD) by weighted gene correlation network analysis (WGCNA), and explore related functions. Methods: GSE66107, GSE93798, GSE30528, GSE32591 gene expression data including podocyte injury models or glomeruli in CKD patients were downloaded from the GEO database. R was used for data analysis. Differentially expressed genes (DEGs) (FDR< 0.05 or |Fold Change|≥1.5) in GSE993395 were assessed by WGCNA. According to Gene Ontology (GO) and known podocyte standard genes (PSGs), podocyte injury-associated modules were defined, with hub genes selected based on average intramodular connectivity. The Cytoscape software was used for network visualization. Nephroseq was used to assess the clinical significance of hub genes. Small interfering RNA (siRNA) was used to evaluate the roles of hub genes in podocyte injury Results: Totally 7957 DEGs were screened, with 15 (co.DEGs) altered in all 4 datasets; 4031 DEGs were used for WGCNA, encompassing 12 modules. Green modules (most PSGs and co.DEGs) were significantly enriched in glomerular development, and considered podocyte injury-associated modules. Furthermore, MAGI2 (a hub gene) was also a co.DEG and PSG. Glomerular MAGI2 levels were reduced in various kidney diseases, and positively and negatively associated with glomerular filtration rate and urinary protein levels in CKD patients. Moreover, MAIG2 knockdown reduced NPHS2, CD2AP and SYNPO levels, and induced podocyte rearrangement and apoptosis. Conclusion: MAGI2 identified by WGCNA regulates cytoskeletal rearrangement in podocytes, with its loss predisposing to proteinuria and CKD.