Hi @khemlalnirmalkar ,
I have moved this post to “other bioinformatics tools” since you are asking how to make a custom box plot not using QIIME 2. There are other software packages you can use for this, e.g., seaborn or plot9 in python, or ggplot2 in R — I recommend checking out their documentation to figure out which works best for you. I cannot offer any specific advice or support on these packages.
Just to note, though:
The barplots generated by
beta-group-significance are provided together with a PERMANOVA test, but these are not “based on” permanova in any way — they simply show the pairwise distances between each sample of group X and group Y. Hence the N increases because it shows the number of comparisons being performed, and is the sample size that should be shown when displaying this type of boxplot.
Sounds like you want that same type of pairwise boxplot — unless if e.g., you have a single reference sample that you want to compare to all other samples. That is the only way that you would have a boxplot where the number of comparisons equals the real sample size for a single group.