Abstract
Acute kidney injury (AKI) is a serious condition characterized by a rapid decline in renal function, leading to severe complications. Recent findings suggest that cuproptosis-related genes (CuRGs) influence AKI mechanisms. This study investigated CuRGs' role in AKI progression and aimed to develop a predictive model for early diagnosis. We utilized the GSE30718 dataset to identify 46 CuRGs, with 16 differentially expressed CuRGs (DECuRGs) found via analysis with ggpubr and pheatmap. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses were conducted using clusterProfiler and Metascape, along with gene set enrichment and weighted gene co-expression network analysis to explore DECuRGs concerning AKI. We also developed machine learning models, including extreme gradient boosting, a generalized linear model, random forest, and a support vector machine, to predict AKI risk. Our analyses found 16 DECuRGs in samples from patients with AKI, which were significantly enriched in α-amino acid catabolism and copper ion homeostasis. A robust brown module of 2906 genes correlated with AKI was established via weighted gene co-expression network analysis. The intersection of module genes and DECuRGs identified 6 hub genes (NFE2L2, DLST, GLS, ATOX1, SF3B1, C6orf136). Machine learning results showed that extreme gradient boosting and random forest models had superior predictive performance, achieving the highest area under the curve values. CuRGs play a crucial role in the pathogenesis of AKI, and the predictive model developed in this study could enhance early diagnosis and guide therapeutic strategies. Future research should validate these biomarkers clinically to help improve patient diagnosis.