Re: How to handle distinct ASVs identifiers assigned to similar taxa to identify enriched or depleted ASVs using DESeq2?

Hello everyone! I used QIIME2-2020.11 for the 16S sequence processing and taxonomy assignment. I performed a differential abundance test using DESeq2 and found the identical genus were enriched and depleted. I scanned the ASVs and realized that certain distinct ASVs identifiers were assigned to similar taxa (see DESeq2_result). I read an interesting topic in the QIIME2 forum with a similar issue here:(What is the reason why there are multiple OTU identifiers in the same species? - Taxonomy assignments in QIIME), however, it couldn’t mention how to deal with such result in downstream analysis.

Hi @muluoljira,

So I think it’s important here to understand the notion of ASVs vs OTUs. I recommend this short-ish paper for a start.
Remember that ASVs are like 100% OTUs, so even a single nucleotide differences between 2 sequences would diverge them as 2 distinct ASVs.

Take your results for example, ASV_1619 and ASV_2862 are not identical sequence-wise but both belong to the genus Allorh… The classifier does not have enough information to recommend a species assignment to either. This is very common with 16S data, we typically don’t have high enough resolution to get species level designation. In other words, we don’t know if these 2 ASVs are of the same species or different, or even how closely related to each other, you can dig deeper into this question using a phylogenetic tree and view where they are in relation to each other.
So you can imagine that 2 species belonging to the same genus can have very different biological functions. In fact, even different strains of the same species can have very different functions/roles in a community. Therefore what you see here, that members of the same genus behaving different is very common and biologically plausible.

If the biological question you have in mind can be answered by say genus level classification alone, you can collapse your data down to the genus level and re-run DESeq2 with the collapsed table. I would generally advise against this but there are certainly some situations for which this makes sense.

Hope this helps a bit.


Thank you so much for the swift response with insightful advice. Its really incredible and helpful!

1 Like