Fig. 3

Here we show the first part of the summarised diagnostics for oHHMed with gamma-poisson emission densities as employed on counts of the average number of protein coding genes along the human genome. Panel A shows the mean (black) and median (grey) log-likelihood of fully converged runs of the algorithm with different numbers of hidden states, with the dashed horizontal line marking the chosen number—which is three. In panel B, boxplots of the posterior (i.e. inferred) mean gene densities of the inference run with three hidden states are presented. Panel C shows the observed distribution of gene counts with the inferred means superimposed as vertical lines. These are significantly different on the 95% confidence level as per one-sided poisson rate test (part of standard oHMMed output). Once again, full descriptions of the diagnostics available for oHMMed can be found in our in our usage recommendations on GitHub [40], and the code for this visual summary plus the corresponding rootogram is available as an R script named “oHMMedOutputAnalyses.R” [39] on GitHub