Figure 4.

The ROC curves of array-CGH data analysis methods. (A) The complete plot. All curves were calculated using the same data set. They show that our Bayesian algorithm (black line) is appreciably more sensitive than all other methods (gray lines) at low (< 10%) false positive rates. (B) Details of the ROC curves in the low FPR region of (A) inside the box with dashed border. See Lai et al. [28] for the identities of the gray ROC curves. TPR, the true positive rate; FPR, the false positive rate.