repeat values for alpha metrics

[tyler1 (tyler)]

hey all!
I believe this is a technical question but correct me if I'm wrong.
I am getting many repeats of values on alpha diversity metrics (shannon, pielou) but not Faith pd.
I hope the picture uploaded, if not, an example would be 20 samples have an evenness of 2, and 15 samples have an evenness of 1.4. this doesn't seem possible, right?
This has only recently started happening on the same data set all though I never changed the code, as old screenshots I have show variable evenness.
I am a qiime newbie but this is really stumping me

[lizgehret]

Hi @tyler1,

Welcome to the forum!

This does look unusual - with that being said, there are no visible groups in the screenshot you provided (i.e. grouping the data by some value in your metadata), so it's hard to see the full context for this visualization.

You mentioned that this has only started happening recently, and that you are using the same dataset - how exactly are you producing this visualization? Has anything else changed since you started running through this analysis (different machine, environment, etc)? Can you provide all of your commands within QIIME 2 that you used to generate the alpha diversity (terminal commands/parameters)? I would recommend that you try running through all of the commands step-by-step, possibly in a new environment (if using QIIME 2 within a conda env) to make sure you are starting completely fresh when troubleshooting this.

Cheers

[colinbrislawn (Colin J Brislawn)]

Hey @tyler1,

Could you post the qza file of alpha diversity values used to make that plot? We want to check out the provenance and see if some sort of filtering could be making these repeated 'bins' of shannon values.