Fig. 5: Impact of the number of markers and the training sample size on the cancer detection performance.
From: Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer
a Performance of using the union of top M cancer-specific markers of four cancer types. Source data are provided as a Source Data file. b Performance of using the union of top M tissue-specific markers of each tissue pair. Source data are provided as a Source Data file. c Performance of the ensemble model for cancer detection increases with increasing training sample size (using 30% to 100% of the training samples). Source data are provided as a Source Data file. In all figures, the points and error bars represent the average AUROC over 10 runs and 95% confidence intervals (\(n:\) 102 test samples per run).
