Identification and validation of key autophagy-related genes in lupus nephritis by bioinformatics and machine learning
Fig 5
Identification of signature genes based on lasso analysis and random forests.
(A) LASSO analysis of cross-validation curves. The optimal λ value was determined using 10-fold cross-validation in the retraining set (left). Penalty score plot of LASSO coefficients for signature genes associated with lupus nephritis in the training set (right). (B) The bubble plot illustrating the relative importance ranking of genes in the random forest model within the training set. (C) The venn diagram of intersection of LASSO analysis and random forest signature genes. (D) The expression levels of five signature genes in individuals with lupus nephritis in the training set compared to normal controls. (E) ROC analysis of five signature genes in the training set.