Fig. 5

a Area under the precision-recall curve (PR-AUC) for each algorithm as a function of increasing guide binding efficiency. Lower guide binding efficiency implies that more guides from essential genes look like the null distribution. The bar heights and error bars are calculated respectively as the mean and standard error from three replicate simulations. b Empirical false discovery rate at an estimated false discovery rate of 0.1. Plotted values are the average over three simulations. Missing values mean that the algorithm did not identify any significant genes at an FDR of 0.1