Integrated bioinformatics and machine learning reveal key genes and immune mechanisms associated with uremia

Uremia is a serious complication of end-stage chronic kidney disease, closely associated with immune imbalance and chronic inflammation. However, its molecular mechanisms remain largely unclear. In this study, we analyzed transcriptomic data from the GSE37171 dataset to identify genes associated with uremia. Differential expression and WGCNA analyses were used to screen core genes, followed by machine learning (LASSO, Random Forest, SVM-RFE) to identify key feature genes. GSEA and immune infiltration analyses were conducted to explore functional pathways and immune relevance. ROC curves were used to evaluate the discriminatory power of the selected genes. Four feature genes-NAF1, SNORD4A, CGB3, and CD3E-were identified. These genes were enriched in pathways related to apoptosis, immune regulation, and oxidative stress. Their expression levels correlated with multiple immune cell types, and ROC analysis demonstrated good discriminatory performance between uremia and healthy samples. Our findings provide potential molecular candidates for further investigation into the immune-related mechanisms of uremia.