Interpretation of alpha rarefaction curves for features

[iptz1 (Ilaria Patuzzi)]

Hi,
I'm running qiime diversity alpha-rarefaction to obtain rarefaction curves from my feature table and I don't know how to interpret my results. In particular, these are the statistics for my table:

This is the rarefaction plot:

And this is the command I am using:

qiime diversity alpha-rarefaction _
--i-table table.qza _
--i-phylogeny rooted-tree.qza _
--p-max-depth 87000 _
--p-steps 50 _
--m-metadata-file metadata.tsv _
--o-visualization alpha-rarefaction.qzv_

My questions are about the y-axis label:

1. is the name "OTU" used as synonym of "single variant"?
2. is it right to say that the curves show, for each sample, how the number of non-zero features grows increasing the sequencing depth?

Thank you for your king attention and suggestions!

[Nicholas_Bokulich]

Hi @iptz1
Thanks for posting!

@iptz1

is the name “OTU” used as synonym of “single variant”?

Yes. The name of that metric is a bit of a hangover from QIIME 1; you can also think of sequence variants as 100% OTUs in a sense (depending on whether you consider the idea of a sequence variant to be intrinsically linked to denoising functionality in a method like dada2) so the name is not entirely inaccurate.

@iptz1

is it right to say that the curves show, for each sample, how the number of non-zero features grows increasing the sequencing depth?

I am not sure I follow 100% but I believe the answer is yes. For each sequencing depth X shown along the X axis, X sequences are randomly drawn from the pool of sequences assigned to each sample. The number of unique features represented by those sequences is calculated. So this would only represent non-zero features if I follow your meaning.

In general, more information on interpretation can be found in this tutorial.

I hope that helps!

[iptz1 (Ilaria Patuzzi)]

Thank you for your help!