Fig. 2: Construction and validation of the ferroptosis-related diagnostic model for IR-AKI.

a Performance curve of the Random Forest model used to evaluate model fitting. b Bar plot showing the top ten ferroptosis-related genes ranked by importance scores in the Random Forest algorithm. c Expression validation of five ferroptosis-related genes identified by Support Vector Machine-Recursive Feature Elimination (SVM-RFE). d Venn diagram illustrating the overlapping feature genes identified by differential expression analysis, Random Forest, and SVM-RFE algorithms. e Forest plot of the multivariate logistic regression analysis demonstrating odds ratios and 95% confidence intervals for the five ferroptosis-related genes; only TFRC showed statistical significance. f Receiver operating characteristic (ROC) curve showing the diagnostic performance of the five-gene model, with area under the curve (AUC) = 0.908. g ROC curve generated from internal bootstrap validation. h Calibration curve comparing the predicted probabilities versus observed outcomes, demonstrating excellent model calibration. i Nomogram constructed based on the five ferroptosis-related genes to estimate individual AKI risk. j Decision curve analysis (DCA) evaluating the net clinical benefit of the nomogram across a range of threshold probabilities. k Schematic representation of the in vivo experimental workflow. l Western blot analysis validating increased protein expression of NGAL and TfR1 in kidney tissues from IR-AKI mice. *p < 0.05; ****p < 0.0001.