Fig. 4

ROC curves for the selection of hit thresholds. The ROC curves were plotted for both error-free hits (a) and 1-error-tolerant hits (b) to select the hit thresholds. Diagonal lines, also known as lines of no discrimination, were plotted as references of random guess. The threshold evaluation was based on a percentage calculated as the number of hits divided by the estimated read depth. A set of percentages was used as the thresholds to calculate the respective true positive rates and false positive rates, which were plotted as the ROC curves. Both ROC curves show constantly high true positive rates and low false positive rates, with the areas under the ROC being 0.9999997 and 0.9999998, respectively