Fig. 5
From: A role for SETD2 loss in tumorigenesis through DNA methylation dysregulation
(A) Multiple random-sampling at different training set sizes shows that as the training-set size increases, the misclassification error greatly reduces. The interactive report shows the best model highlighted as the red circle at training set size of 98. (B) Heatmap displaying the methylation of the top 3 CpGs (cg14297023, cg17054691, and cg25415966) selected as the best model. Columns clustered using complete-linkage clustering and distance measure given by the Euclidean distance. (C) PCA analysis using the best model shows a good separation between SETD2 cases and WT renal cancer samples. (D) ROC curve to display the true positive and false positive rates of the signature in classifying SETD2 controls and cases.