Tracing the molecular landscape of diabetic nephropathy: Insights from machine learning and experiment verification.

OBJECTIVE: Diabetes is a chronic disease resulting from insufficient insulin secretion or impaired insulin function. Diabetic nephropathy (DN) is one of the most common complications of diabetes and a leading cause of end-stage renal disease. Early diagnosis of DN is crucial for timely intervention and effective disease management. METHODS: Gene expression profiles GSE142025 and GSE220226 were retrieved from the GEO database and combined into a metadata cohort, while GSE189007 was obtained as an independent validation dataset. Differentially expressed genes (DEGs) were identified in 46 glomerular samples from DN patients and 31 control samples. Gene Ontology (GO) and Disease Ontology (DO) enrichment analyses, gene set enrichment analysis (GSEA), least absolute shrinkage and selection operator (LASSO) regression, support vector machine-recursive feature elimination (SVM-RFE) analysis, and area under the curve (AUC) calculations were performed. RESULTS: A total of 109 DEGs were identified. Among them, DUSP1, EGR1, FPR1, G6PC, GDF15, LOX, LPL, PRKAR2B, PTGDS, and TPPP3 were selected as potential diagnostic biomarkers for DN. These biomarkers exhibited a positive correlation with immune cell infiltration. Experimental validation identified LOX as the most promising novel diagnostic biomarker for DN. This study provides new insights into the early diagnosis, pathogenesis, and molecular mechanisms of DN.