I am attempting to create a series of grouped barplots from filtered feature-tables which each include only one of three tissue types. For example, I have three different tissue types A, B, and C, all of which were subjected to the same three treatments - 1, 2, and 3. There are numerous other metadata that I would like to ignore for the purposes of these barcharts (keeping all of them in the data IOW). I have used the filter-samples/p-where command to filter the initial table to keep only tissue type A. I would like to create grouped barplots of the effects of all three treatments (1,2,and 3) on OTU composition of sample A without regard to any other metadata. However, when using the filtered feature-tables, the barplots keep separating everything by the other metadata (like collection date, etc.), even though I am specifying “treatment” in the --m-metadata-column portion of the qiime feature-table group command. Any ideas as to why this is happening?
I have absolutely no problems creating these grouped barplots using the unfiltered feature-tables from the analysis with this very same metadata table. What am I doing wrong?
OK, here are the filtered .qza file and the resulting .qzv files. As you can see in the .qzv file, the plot is keeping every sample separate by treatment (UW, W, WW), by rep (1-8), and by sampling time. I am trying to group all of this so that I see only calyx samples grouped together by the three treatments so that there are only three columns (UW, W, WW). Thanks!
@eburchard,
The provenance on this file does not indicate that you used feature-table group. So far this barplot makes sense based on what you did — filtered to only calyx samples. Use group on the QZA file you attached, and then build a barplot to get what you want:
You mentioned above, though, that this did not work:
If this is still not working, could you please share that barplot and feature table?