sample not in taxa bar plot

Even though the sample has a feature count of 3037, it is not shown in taxa barplot. I have tried close reference with NCBI sequence, and then tried with silva classifier but the result is not shown in taxa barplot for both ways.

Analysis includes 6 samples and for the other 5 samples, results satisfy the expectation. I’m pretty confused about this. Could someone help me?

Thanks…

Hi!
Could you attache your metadata and barplot.qzv?

PS Just in case - sometimes there is a scroll bar under the barplot and some samples are not shown so you need to scroll it to see it.

Here is metadata: metafile.tsv (226 Bytes)
Here is barplot: taxa-bar-plots.qzv (360.3 KB)
Here is table after dada2: table.qzv (419.9 KB)

I haven’t seen any scroll bars since this is an analysis for 6 samples. I was expecting problems with OK-6 because it was from a different run but I got problem with OK1.

taxa-bar-plots.qzv (733.7 KB)
table.qzv (410.1 KB)

These are the results of the analysis done with only forward read. Still, OK1 doesn’t show up.

The table summaries and taxa barplots are not matching up, i.e., the summaries are not of the tables that were actually used to generate the barplots.

Check the provenance to make sure you are using the correct files. And compare the table summaries before and after using filter_features.

I suspect the problem here is that the sample is being dropped from the feature table when you use filter_features because no features are left in that sample (e.g., because all are chloroplast).

3 Likes

@Nicholas_Bokulich
You did it! Thank you…

They were filtered out because they are chloroplast and mitochondria features, and silva classifier had them.

I’m not sure if I should post a new topic about this but, is it wise to filter out these features? Are they supposed to be amongst 16S sequences? Is there any material that you know of I could read about this?

Chloroplast and mitochondria? Definitely filter them out unless if you want them there. These are usually host DNA or in any case usually not microbial so are just “noise” on your plots and can massively skew your data. By way of example I will tell you a story. :open_book:

Once upon a time I was studying the microbiota of grapes from different sites. :grapes: :petri_dish:

Some grape samples really stood out on a PCoA plot and I thought “wow, what makes those grapes so special?” :thinking:

…nothing, it turns out, just an inordinately high amount of host DNA. Filtering that out made the profiles comparable to everything else once again. :tada:

1 Like

Thank you very much.

Now, I’m pretty sure there is a problem with the treatment of the sample.

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