Fig. 6

The Random Forest construction and receiver operating characteristic (ROC) curve of candidate genes. a The flow chart of Random Forest construction. ROC curve for the relative expression of HCC (n = 84) and non-tumor (n = 11) mRNA-seq samples of each validated gene and gene combination. The corresponding area under the curve (AUC) value is indicated. Diagonal lines represent the performance of a random classifier. b ROC curves of three-gene sets for classifying non-tumor samples from HCC samples in the TCGA test set. All three-gene sets achieved an AUC > 0.82. c ROC curves of two-gene sets for classifying non-tumor samples from HCC samples in the TCGA test set. All two-gene sets achieved an AUC > 0.65. d ROC curves of 6 candidate genes for classifying non-tumor samples from HCC samples in the TCGA test set. All 6 genes achieved an AUC > 0.6