Skip to main content
Fig. 4 | Genome Biology

Fig. 4

From: LABS: linear amplification-based bisulfite sequencing for ultrasensitive cancer detection from cell-free DNA

Fig. 4

Integrating multiple layers of information from the LABS provides a better prediction. a Diagram of the integrated model. TSS methylation and copy number variations were first analyzed by PCA to find the most informative PCs. b ROC curves showing better prediction accuracy of the integrated random forest model for CRC compared to methylation biomarkers alone. CNV, copy number variation; ROC curve, receiver operating characteristic curve. c Variable importance of the random forest model shown in b. Each bar represents a variable, i.e., a principal component based on TSS methylation/a copy number variant, or the relative percentage of a particular immune cell type. d ROC curves showing better prediction accuracy of the integrated SVM model for CRC compared to methylation biomarkers alone. SVM, support vector machine

Back to article page