Fig. 4.

5hmC markers’ prediction for treatment response in the training and validation cohort. a, b Heatmaps of 13 5hmC markers with treatment response, batch and sex information labeled in the training and validation cohorts. Unsupervised hierarchical clustering was performed across genes and samples. c Receiver operating characteristic (ROC) curve of the classification model with 13 5hmC markers in the training and validation cohorts. The true-positive rate (sensitivity) is plotted in function of the false-positive rate (1-specificity). d Confusion matrix that shows the model performance in the validation cohort (responders: 22, non-responders:8). e ROC curve of the classification model with LDH, stage and LDH combined with stage for DLBCL patients