Related Experiment Video
Updated: Jan 12, 2026
Supervised Machine Learning for Semi-Quantification of Extracellular DNA in Glomerulonephritis
Published on: June 18, 2020
Construction and validation of a hypertensive nephropathy diagnostic model based on mitochondrial-related genes using
Gonglin Tang1, Guixin Ding1, Li Xie2
1Department of Urology, The Affiliated Yantai Yuhuangding Hospital of Qingdao University, NO. 20 East Yuhuangding Road, Yantai, 264000, Shandong, China.
Background:
Hypertensive nephropathy (HTN) arises from chronic hypertension and may potentially result in severe renal failure. Due to the absence of reliable and efficient biomarkers, the diagnosis of HTN is constrained. Mitochondrial-associated genes are closely associated with the pathogenesis of HTN. Our study aims to find new reliable diagnostic and therapeutic targets for HTN.
Methods:
We obtained the dataset GSE37460 from the GEO database and performed Weighted Gene Co-expression Network Analysis (WGCNA) to analyze the expression modules of differentially expressed genes (DEGs) identified from GSE37460, in conjunction with mitochondrial-related genes (MRGs) for hub gene selection. We employed KEGG, GO, and GSEA enrichment analyses to elucidate the differential genes. Subsequently, diagnostic models were developed using LASSO, SVM, and RF algorithms. Furthermore, functional analysis between the disease group and control group in GSE37460 was conducted using ssGSEA, and the diagnostic performance was evaluated through Receiver Operating Characteristic (ROC) curve manipulation.
Results:
Through differential analysis and WGCNA analysis, we have identified a total of 695 DEGs. Cross-analysis revealed that among these genes, 14 are MRGs. These genes are primarily enriched in biological processes and pathways related to NOD-like receptor signaling, coronavirus disease-COVID-19, necroptosis, tumor necrosis factor production, nuclear envelope, transmembrane transporter binding, and protein tyrosine kinase activity. Furthermore, utilizing LASSO, SVM-REF, and Random Forest algorithms, we constructed a diagnostic model consisting of three genes (PYCARD, NRP1, and IFI27). The resulting ROC curve demonstrated high diagnostic accuracy, with an area under the curve (AUC) value of 0.998. In addition, ssGSEA analysis revealed significant correlations between these three hub genes and heme metabolism, inflammatory response, cholesterol homeostasis, interferon response, Wnt/β-catenin pathway, and epithelial mesenchymal transition.
Conclusion:
We have elucidated the role of MRGs in the pathogenesis of HTN and developed a diagnostic model with high diagnostic potential.
