Fig. 11

Variability of the overlap between the PDFs of the two distributions described in Fig. 10d. The panels show histograms of \(N=1000\) simulated overlap values which are simulated as described in the main text. Left: We assume that the estimates of the orange distribution relied on 60 single cells and the blue distribution on 200 single cells. Right: For both distributions, parameters are assumed to be estimated on 60 single cells. The 86% overlap of the original PDFs from Fig. 10d, i.e. \({\mathcal{L}}{\mathcal{N}}({\hat{\mu }}_{1,cases}=2.10,{\hat{\sigma }}^2_{cases}=0.19^2)\,\text{and}\,{\mathcal{L}}{\mathcal{N}}({\hat{\mu }}_{1,controls}=2.03,{\hat{\sigma }}_{controls}^2=0.20^2)\), is marked in turquoise. The light grey bars of the histogram indicate values below the empirical 5%-quantile. If the original overlap falls into this range, we reject the null hypothesis that both distributions identical