High yield of D_2__Alphaproteobacteria;D_3__Rickettsiales;D_4__Mitochondria in samples from wild bats... Artifact?

Hi!,
I'm working on the microbiome of a wild nectar feeding bat. This samples may contain a lot of plant material including Pollen. I was using the V4 region (515-806 primers) and extractation from fecal pellets.
Before I ran all my samples, I decided to run 15 samples and one control jsut to check if everything went ok.
The output of the run was quite particular and I found that a large percentage of the microbiome in 3 samples matches the following string:

"D_0__Bacteria;D_1__Proteobacteria;D_2__Alphaproteobacteria;D_3__Rickettsiales;D_4__Mitochondria"
(see image)

I know that it is normal to get some mitochondria in the MB because the primers link this too. The problem is that this didn't went away with the filtering step:

"qiime taxa filter-table
--i-table output2/table-dada2.qza
--i-taxonomy output2/taxonomy.qza
--p-exclude Archaea,Mitochondra,Chloroplast,Unassigned
--p-include D__
--o-filtered-table output2/table-filtered.qza
"
There is one specific ASV that account for 10% of all the reads in my run!! and this is mainly concentrated in 3 samples. This seams wrong, so I explored different options to solve this, first I evaluated if this was a mistake in the bioinformatic processing:

In trying to correct this I:
+Re-ran the QIIME2 pipeline again with a less conservative cut 215 instead of 200 in the DADA2 step
+re assigned the taxonomy using Silva (the first try was with greengenes)
Conclusion: it's not something in the pipeline as far as I can tell (or is it?)

The second though was that this could be DNA from the Host, I Aligned the problem ASV, a "normal" ASV from my MB, a 16S sequence from my host and the sequence of the 2 top hits after blasting the problem ASV. The results, the problem ASV and the 16s sequence from the host do not aligned at all. Also the ASV and the sequence of the top hits are perfect matches while the "normal" ASV is quite different from the others.

When I blasted the problem ASV, the result the top hits are Mitochodrial DNA from plants that are NOT present in my study area. I'm starting to think that this sequence may come from undigested pollen that just when through the bat directly.

I was wondering if anyone was ever had a problem like this one... Im curious if this is a failure in the filtering step or if there may be some problem with the samples.
Any input would be greatly appreciated!

Luis

Thanks for the detailed troubleshooting description!

I agree, this is most likely pollen mitochondrial DNA (host was my first guess, but you solved that one).

Top-hit BLAST will not necessarily be all that informative, i.e., for species-level identification of very short mitochondrial 16S sequences. I think it is useful in that it has indicated that this is plant, not host DNA. But the fact that the top hit is to a plant not in the study area is probably not too important — because the mito 16S for related species that are in the study area are probably quite similar, leading to this misclassification. You could BLAST against a database containing only plant species found in the area, but that may be an exhaustive list...

All in all, this sounds like an expected result, BLAST is just giving you the wrong ID.

Thanks a lot for the reply!
I also think this could be Pollen. My concern here is that the mit DNA from the pollen could be out competing the bacterial DNA due to the high concentration of it in the fecal pellets. Due to this high abundance, if I remove them from the samples by extra filtering steps a lose a lot of reads (up to 92%) in those samples. I have some pollen sample from that plant here, so I’m extracting using the same protocol and will do a new test to see if this is the source.
Again thanks for the reply and I’ll update this when I have further information.
cheers!

1 Like

That appears to be the case here, but unfortunately there is nothing you can do about that with the current data — you will still need to filter and accept that you lose several samples in the process. You could try to find different primers that do not amplify plastid or mito 16S, but otherwise you are stuck.

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