run_otsu_discordance()
In the above plot we observe that not only are the density curves for each marker in the analysis far better aligned after normalization, we also see that the Otsu thresholds (ticks on the x-axis) have moved far closer than in the raw data. In general, also note that all plots generated using mxnorm are ggplot2 plots and can be adjusted and adapted as needed given the ggplot2 framework. We can also visualize the results of the threshold discordance analysis stratified by slide and marker, with slide means indicated by the white diamonds:
Note that for each slide and marker pair in the dataset (denoted as colored points in the above plot), we see a reduction in threshold discordance in the normalized data compared to the raw data. Further, we also see dramatic improvements in the mean threshold discordance denoted by the white diamonds for the normalized data.