Fig. 3

Distribution of estimates from IUPMStats and IUPMBayes for simulations given IUPM = 1. We used a barplot to summarize the distribution of IUPMStats estimates, which makes clear that these estimates are limited to a relatively small number of values given the dilution series and number of replicate wells in our simulation experiments. The leftmost bar (red) corresponds to the median Bayesian posterior estimate employed by IUPMStats when all wells are negative. The distributions of estimates from IUPMBayes under two sets of variant frequencies (1:1 and 1:1:1:1:1) are summarized with Gaussian kernel densities (curves). Unlike the IUPMStats estimates, these distributions were unimodal and centred near the true value. We obtained similar results under varying conditions and IUPM values