How to interprete box plots beta diversity plot ?

User Support
1

Hello,

I have to compare beta diversity between 2 groups , sars cov positive ( n = 20) versus sars cov negative group (n = 20).

when using permanova test , i generated box plots and i got in x axis (group) n = 190 for negative group and n = 400 for positive group !

What do 190 and 400 numbers refer to ?

2

Would you be willing to share the .qzv files that contain the results of the statistical tests?

I have a hunch about these numbers, but let's confirm by looking at the results first.

3

Hello
@colinbrislawn

Please find attached a screen shot of beta diversity.
y axis represents distance, x axis group. Noting that the sample size is 40 (n= 20 negative versus n = 20 positive). Please indicate What do 190 and 400 numbers refer to.

beta
beta591×442 7.4 KB

4

Alpha diversity measure diversity inside a single sample:
Sample1: 10 :bird:
Sample2: 4 :bird:
Sample3: 10 :bird:
Sample4: 0 :crying_cat_face:

Beta diversity measure diversity across a pair of samples
Sample1&2: 4 :bird: in both samples
Sample1&3: 9 :bird: in both samples
Sample1&4: 0 in both
Sample2&3: 4 :bird: in both
Sample2&4: 0 in both
Sample3&4: 0 in both

Note that for alpha diversity have 3 samples and 3 values (n = 3)
But for beta diversity, we have 6 values because we are comparing all pairs of samples.

Now that you know beta diversity operates on pairs of samples, do you see the pattern?
Hint 1: List out all pairs of samples when n = 5 and n = 6
Hint 2: The formula uses a triangle number. Edit: NOT a factorial
Hint 3: The formula for pairs of samples changes when you are comparing across groups of samples.

PS. Are you able to share the .qza or .qzv file?

6

@colinbrislawn

I already shared a screen shot of beta diversity in box plots.

Please indicate What do 190 and 400 numbers refer to.
Thanks.

7

Hi @M_F

These are the numbers of pairwise comparisons represented by each distribution (box) in the boxplot.

The plot shows the distances to the negative control samples.
The positive vs. negative comparison is easy: 20 pos. X 20 neg. = 400 pairwise comparisons
The negative vs. negative would be the total number of pairwise comparisons between each negative sample and each other negative sample (no self comparisons) so if you have 20 samples in the negative group, the number of distances is not simply 20 X 20, but 19 + 18 + 17 + 16 + 15 + 14 + 13 + 12 + 11 + 10 + 9 + 8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 = 190 pairwise comparisons within the negative group

1 Like
8

Thanks so much @Nicholas_Bokulich