Top: As in Fig 2A. Additional examples of variability in effect size after binning into different numbers of internal replicates (4 internal replicates, left; 16 internal replicates, center; 64 internal replicates, right). Bottom: Boxplot showing the distribution of the variability across the whole data set for 64 internal replicates. The median absolute deviation in effect size (log2 fold change) is plotted for all guides in the library. Note that the spread is substantial, this is the case whether or not RSLs are present, but can only be detected in the presence of RSLs. Box, first and third quartiles; whiskers, 1.5‐fold interquartile range; line, median; cross, average. Outliers are shown.