Skip to main content
Fig. 4 | Genome Biology

Fig. 4

From: MAUDE: inferring expression changes in sorting-based CRISPR screens

Fig. 4

MAUDE simulations propose optimal cell-sorting bin configurations. a Evaluation of expression bin configurations by simulation. Left: an experiment is simulated (10 replicates, top), followed by selection of each bin configuration, MAUDE analysis, and evaluation of performance for true positive recovery and effect size prediction (by Pearson’s r). Right: each bin configuration had 2–6 bins that are contiguous on each end and symmetric about the center, progressing gradually with inclusion from outer to innermost: A and F—outermost and always included, B and E (when present)—contiguous with A and F (respectively) and more central, and C and D (when present)—contiguous with B and E (respectively) and innermost. b, c Identification of optimal bin configurations. Performance (y axis) of each bin configuration (x axis) for b Pearson’s r between the true effect sizes (as defined in the simulation) and the predicted effect sizes (by MAUDE), and c the number of true positives (elements defined as having an effect in the simulation that are predicted to have an effect by the analysis). x axis labels: bin sizes expressed as the percent of the distribution covered by each of the paired corresponding bins, sorted in order of increasing performance. Error bars: standard error of the mean (SEM) for the 10 simulated experiments. Gray shading: number of bins in the configuration

Back to article page