Screening for potential biological markers of chronic kidney disease based on WGCNA and machine learning

Background: Chronic kidney disease (CKD) is defined as persistent urinary tract abnormalities, structural abnormalities, or impaired excretory function of the kidneys, which is the 16th leading cause of years of life lost worldwide and places an enormous burden on medical care. However, the mechanisms for the progression of CKD are still poorly understood. Therefore, the aim of this study is to identify the genes responsible for CKD and to establish a genetic diagnosis model. Results: : By combining the differentially expressed genes with the Weighted correlation network analysis, a total of 264 differentially expressed genes, most associated with CKD were identified. According to the results of GO enrichment analysis, we confirmed the molecular functions were most closely related to haptoglobin binding and oxygen carrier. In the biological process, the term most related to oxygen transport, and in cellular components was hemoglobin complexes. KEGG enrichment analysis showed that these DEGs were related to pancreatic secretion, protein digestion, and absorption. Subsequently, ZCCHC7, ZNF396, and EIF4E3 were identified as three potential biological markers in the kidney of patients with CKD by using Least absolute shrinkage and selection operator (LASSO) regression and machine learning algorithms. Gene Set Enrichment Analysis (GSEA) furtherly demonstrated that three genes respectively involved in lipid (ZCCHC7), glucose (ZNF396), and metabolism (EIF4E3).A diagnostic model was also constructed based on these three genes, and the ROC curve showed that the 3-gene diagnostic model has a good fit. Finally, the three potential biological markers of CKD and their model were validated by using GSE175759, and the results further indicated the diagnostic value of these three potential biological markers. Conclusions: : We successfully identified and validated that ZCCHC7, ZNF396, and EIF4E3 are potential biological markers in CKD patients, which may influence the progression of CKD via the metabolism of fat, sugar, and steroid hormones. Our findings offered a potential diagnostic biomarker for CKD.