Sub-sampling depth not the reflected when downloading taxa barplot data


I am pretty new to QIIME and coding etc. so please forgive me if this is a silly question.
I have followed through the tutorials and have been using QIIME2 to look at my 16S data.

When I am up to the alpha and beta diversities script and have to choose my sub-sampling depth I run the following:

qiime diversity core-metrics-phylogenetic
–i-phylogeny rooted-tree.qza
–i-table table.qza
–p-sampling-depth 4000
–m-metadata-file sample-metadata.tsv
–output-dir core-metrics-results

Later down the pipeline when I am creating the taxa barplot and downloading the output, I notice that the samples with no reads or below the sub-sampling thresholds are still there and the other samples which I expect to be sub-sampled at 4000, aren’t and are showing their full amount of reads.

Am I missing a step somewhere?

Thank you

1 Like

Hi @ChloeD93!

No, you aren’t missing a step - I would recommend taking a closer look at the command you are running. I say this because we recommend using the unrarefied feature table for this taxa barplot visualization - rarefying a feature table is only important for diversity metrics that are sensitive to the sampling depth - taxa barplots do not fall in this category. So, when you are running you qiime taxa barplots command, you are providing your unrarefied table, not the rarefied one. Make sense? Let us know if you have any more questions! :t_rex:

1 Like

Hello Matthew,

thank you for your prompt reply. I just wanted to clarify something, when it comes to doing analysis on the samples (in PRIMER for example), like QIIME1, they should be subsampled to the chosen depth. In this case, would you use your rarefied table in the barplot command and then download that CSV file to do the analysis on? or is there a different command to get your subsampled data (rarefied) without having to make the barplot first?

Hi @ChloeD93, I am not entirely sure I understand your question. Some statistics require an even sampling depth (e.g. a rarefied table), while others don’t. As well, when running an analysis, it isn’t necessarily one-dimensional, where every output is fed directly into the next input - analyses take a few different branches, corresponding to different questions to ask of the data.

I’m not quite sure what you are trying to accomplish here, but you would not want to ever perform this kind of taxonomic analysis using a rarefied table, in my opinion. You are potentially throwing data away, for no reason (the bar plots are just showing the relative abundance, which is not impacted by the number of sequences/features observed in a particular sample, because it is a ratio).

If you are following something like the Moving Pictures tutorial, you will generate a rarefied table when running your diversity metrics (e.g., qiime diversity core-metrics-phylogenetic). There is also a command to rarefy a table, separate of that (qiime feature-table rarefy).

BUT, before you proceed, I would strongly suggest you take a step back from performing your own analysis and focus on understanding the basics - we have a lot of great tutorials at that answer the questions that you are asking here. Thanks! :t_rex:

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