How to interpret beta-group-significance results (boxplots)

[Matilda_H-D]

Hi there,

I was wondering if someone would be happy to explain how to interpret the output of beta-group-significance, particularly the pairwise boxplots.

One thing that's really confusing me is the 'n' numbers under the boxplots (see below for example). Initially I assumed that it referred to the number of samples in that category, but that can't be possible as n=216 for 'supra' when I know there are only 24 'supra' samples and only 35 samples total in my dataset. Is 'n' the number of 'distances' measured? i.e. there were 216 distances between 'sub' and 'supra' samples?

Many thanks for any help!
Matilda

[Nicholas_Bokulich]

Hi @Matilda_H-D,
Yes! N = the number of distances measured. So you have 24 supra, 9 sub, and 1 supra_sub.
Sub total N = sum of all integers ≤ 9 = 36
Supra total N = 24 * 9 = 216
Supra_sub total N = 1 * 9 = 9

One boxplot is output for each group G; the distances between each sample in group G and each other sample are shown. So in the plot you shared above, each box and whisker is showing the distribution of distances between each sample in group “sub” and each other sample.

I hope that clarifies!

[Matilda_H-D]

Thanks @Nicholas_Bokulich!

[Maggi_Mars]

Hi! I followed this thread because I had the same question. Thank you so much for you answer! Unfortunately, I need a little more help.

I have three treatments - CD, NSD, and HSD. The actual number of samples in CD = 6, NSD =8 and HSD =6. In the Distances to CD box plot, CD(n=15), HSD (n=36) and NSD(n=48). I believe the HSD n=36 is 6 HSD * 6 CD and likewise the NSD n=48 is 6 CD * 8 NSD. What I can’t figure out is the n=15 for CD. In your example you wrote 9!=36 but the way I know !, 9! is not 36, it is 362880, so I am not sure how to use the same equation to understand my results. Why is label under CD(6 samples) in the Distances to CD boxplot n=15?

Thank you so much in advance!

[Nicholas_Bokulich]

Hi @Maggi_Mars,
You are correct — I did not mean a factorial (as 9! would denote), I meant the sum of all positive integers less than or equal to n.

If you have 6 samples in a group, the number of distances between all samples within that group will be:
5 + 4 + 3 + 2 + 1 = 15

and so you have n=15 for CD, and @Matilda_H-D has 36 within-group distances from a group of 9 samples.

I hope that helps!

[Maggi_Mars]

Awesome! Thank you so much!