Rarefaction for Beta Diversity Analysis

I am just beginning to learn Qiime2, and I have a question regarding rarefaction.
My understanding is that value chosen for the –p-sampling-depth in the command qiime diversity core-metrics-phylogenetic causes each sample to have that count of sequences.
I do not wish to do that. Is there a way around this rarefaction step in Qiime2?


Just out of curiosity — are you using a different normalization strategy?

The core-metrics-* actions are just pipelines, i.e., they run a series of actions in a single command. So you are able to replicate the methods used in the pipeline with other actions, giving you more control over the individual parameters.

In your case, you can replicate this pipeline without even sampling. So you can do something like run qiime diversity beta to get a distance matrix without even subsampling (that action does not have a sampling depth parameter and in the pipeline rarefying is done prior to that action), then qiime diversity pcoa to generate PCs, then qiime emperor plot to make an emperor plot. It’s more work, but more control.

See the overview tutorial for a flowchart showing what’s going on under the hood in this pipeline — that will let you rebuild these steps without rarefying.

I hope that helps!

1 Like

Thank you, Nicholas!
That was extremely helpful!

I have a few follow up questions:

  1. Could you explain in more detail at which point I would run qiime diversity beta as well as the other command you listed? I was a little confused when you said the “rarefying is done prior to that action.”
  2. Could you possibly include the parameters required for running qiime diversity beta?


See the tutorial I linked to above. You would just run qiime diversity beta on your feature table, then proceed as shown in the flowchart.

type qiime diversity beta --help to see all options, or check out the docs

I hope that helps!

Again, extremely helpful!
I really appreciate it!


1 Like

Hello again,
I have been trying to run the qiime diversity beta pipeline as you suggested. I have successfully got it to run.
However, I am unsure how to use the distance matrices generated to later generate PCoAs.

Using qiime diversity betaI would like have many of the same files that are generated in qiime diversity core-metrics-phylogenetic however the tutorials do not explain how do so in a detailed manner. They only explain how to use qiime diversity core-metrics-phylogenetic.

If you could shed some more light on the qiime diversity beta pipeline, that would be wonderful.

the flowchart shown in the overview tutorial that I linked to above lists all the steps. I am re-linking here for convenience. For each individual command listed in that flowchart, you can check out the plugin docs for usage documentation. Or get the help docs in the terminal by typing the --help flag after a command. E.g., qiime diversity pcoa --help.

For example:

we see that qiime diversity beta produces a DistanceMatrix, which we use as input to qiime diversity pcoa to produce PCoAResults. Check out the help docs for more details on using that particular command.

I hope that helps clarify!

That helped significantly!
Thank you!

Final question, using qiime diversity beta is there a way to generate a table with the OTUs and the Shannon Index Values?

Check out qiime diversity alpha.

Note that alpha diversity analyses really require some type of normalization to account for different sequencing depths, or else your results will be invalid — this is why rarefaction (as the simplest method of sampling depth normalization) is baked in to the core-metrics pipelines.

I assume you know this and are bypassing rarefaction in favor of a different normalization scheme, but just want to pass a word of warning before proceeding.

1 Like

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