Can I perform ANCOM with 2 samples?

[jose_gacia (Jomagrax)]

Hi everyone!

Im using qiime2-2020.8 on conda.

Im trying to perform ANCOM analysis to compare two samples, from high and low salinity. When I try to do so, I get the following error.

qiime composition ancom
–i-table 3.4-7.7/comp-k-3.4-7.7-18S-table.qza
–m-metadata-file …/Metadata.tsv
–m-metadata-column Salinity
–o-visualization 3.4-7.7/ancom-k.qzv

Plugin error from composition:

All values in grouping are unique. This method cannot operate on a grouping vector with only unique values (e.g., there are no ‘within’ variance because each group of samples contains only a single sample).

From this I understand that perhaps you can’t peform this analysis with only two samples.

Is my assumption correct?

[vheidrich (Vitor Heidrich)]

Hi Jose,

You’re exactly right! As most (if not all) statistical tests, ANCOM has no statistical power with only one sample per group. If you want to visualize differences between groups, I would recommend simply plotting the taxa relative abundance per sample. This will not give you conviction about your results, but it sure is a good descriptive tool for this limited dataset.

Good luck,

Vitor