Problems with alpha diversity analysis

I have mice (feces) 16S data from 4 different lines of mice (3 males and 3 females per line). I have finished until feature table creation and bar plots. In the metadata file, I have sex, how long they survived the infection, and strain name (BALBc, C57BL6, etc., ).

  1. I am trying to see if there is any difference within individuals of the same strain
  • I assumed this could be done alpha diversity and tried to do faith’s phylogenetic diversity, shanon indeed and chao1 index.
  • I can only see the difference between males and females (p and q-values), but never as individuals. This is the error I see in .qzv file.

The following metadata columns have been omitted because they didn’t contain categorical data: Age_Feces, BarcodeSequence, Days_survived, LinkerPrimerSequence, SampleCode

The following categorical metadata columns have been omitted because the number of groups was equal to the number of samples, there was only a single group, or the column consisted only of missing data: Description, Good_strain_Name, Location,Resistant, S_R_T, Sam_name, Susceptible, Tolerant****strong text

  • Any way to compare the individuals within the strain?
1 Like

Hi @Aravindh_Nagarajan,

First, welcome to the :qiime2: forum!

You can't do a statistical test with only one sample per group. Statistical tests compare distributions and one observation does not a distribution make. You could, however, ask whether mice within one group are more similar by looking at variance. You could try an f-test for your alpha diveristy, although I don't think it's implemented in QIIME, so you might need to go through the python API or qiime2R or the standard export and into your favorite statistical pacakge.

You could also use beta diversity (pick your favorite metric[s]). I would use permanova/adnois first to see if there is a significant difference between the groups, and follow up with permdisp to see if there's variance. There's some discussion of this in the PD mouse tutorial, as well as around the forum.

Best,
Justine

1 Like

Thank you! the PD mouse tutorial seems interesting, I will try that. Is there a place to learn more about the Python API or qimme2R workflow?

1 Like

There’s a bunch of information about qiime2R in the community tutorial section on the forum. It’s an excellent resource for all sorts of packages and I recommend exploring it.

You can learn more about the python API at the interface section of the QIIME 2 website.

Best,
Justine

2 Likes

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.