Assistance with Alpha Rarefaction curve

Hello everyone!

I made an alpha rarefaction curve. However, it does not have an "All samples" curve where it shows every sample individually making up the curve. Right now my curve is subdivided into the categories in my metadata file.

This is the code I ran:

qiime diversity alpha-rarefaction \

--i-table rarefied-table.qza \

--i-phylogeny rooted-tree.qza \

--p-max-depth 15000 \

--m-metadata-file metadata_file.tsv \

--o-visualization alpha-rarefactionoral100rare.qzv

Would really appreciate some help with this!

Hi @Coinenthusiast ,

Often a metadata file might contain information that is unique for each participant, e.g., in the moving pictures tutorial alpha rarefaction example from the amplicon-docs tutorials barcode-sequence is unique for each sample and can be used as a proxy for per-sample information.

If your sample metadata does not already have such a column, you can create one, e.g., by creating a new column that contains the sample IDs (or however you would like the samples listed in the legend).

Good luck!

1 Like

You’re right! I just made an extra column containing the sample IDs as categorical data. Thank you! I do have another question though. I wanted to try to decrease the intervals between the error bars on my alpha rarefaction plot but got the following error:

client_loop: send disconnect: Broken pipe

Here was the code I ran:

qiime diversity alpha-rarefaction \

--i-table table.qza \

--i-phylogeny rooted-tree.qza \

--p-max-depth 100000 \

--p-min-depth 6200 \

--m-metadata-file PD_T2D_metadata_v2.tsv \

--p-steps 94 \

--o-visualization alpha-rarefactionoral100wminnewsteps.qzv

I am wondering if the issue was that there were too many steps and any help would be greatly appreciated!

Hi @Coinenthusiast ,
Broken pipe normally implies that a job was killed, e.g., due to an out of memory error or time limitation on a cluster. So indirectly yes this could be related to the large number of steps. I do not normally think of alpha-rarefaction as a particularly resource-intensive process, but it looks like you do have a rather high depth and number of steps, so this could be the issue. I suggest reducing the steps and requesting more resources on your cluster.

Good luck!

1 Like

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