top taxa vs top phyla/genera/species; some insights?

[MarwaTawfik]

Hello

Thanks for all of the help you provide. Really thankful.

I worked on my feature table in R for visualisations.
I can see a difference when I visualise after filtering (outside qiime2) between top taxa and top certain taxa (ex: genera).
So I was wondering why is this happening?
In the case of top taxa, I use the whole table subset of the top 10 taxa while in the case of the top genus, I aggregate or glom the taxa to a certain level (genus in this case) and then subset and plot. Also, I do the same for top species, where I aggregate or glom the taxa to a certain level (species in this case) and then subset and plot.

here are some examples of the plots
So I can see that there is a top genus; Flavobacterium

But it is not there in the top species

Do you think using top CERTAIN TAXA is fine?

Many thanks
M

[SoilRotifer (Mike Robeson)]

Hi @MarwaTawfik,

This might be related to the issue discussed in this thread:

Mismatch between qiime2 taxa bar plot and phyloseq taxa bar plot

Hi @Sam_Degregori, We are missing some information on the QIIME 2 side, prior to making the taxonomy plot: What filtering was performed? Were the data rarefied ? Were unwanted taxa removed? That is, did you perform the same operations in QIIME 2 as you did in phyloseq? If not, you are truely comparing apples to oranges here. Within R / phyloseq there are several operations that may be contributing to the observed differences: You've: rarefied the data : rarefy_even_depth(im) ran tax_glom(…

-Mike

[MarwaTawfik]

Many thanks for sharing the post, very good to know.
It is not quite related as my question is what is really different between top taxa (all levels) and top phyla let's say

[SoilRotifer (Mike Robeson)]

@MarwaTawfik

It is not quite related as my question is what is really different between top taxa (all levels) and top phyla let's say

Not necessarily true. If there is no species label for Flavobacterium. Then it will be dropped when you filter at the species level, because the species label is NA. Which may be the reason why it does not appear in the species plot.

Alternatively... many species of Flavobacterium are combined together when grouping at the genus-level... but individually each member species is not enough to register at the top for the species level. Thus, looking at top species will effectively resort compared to top genera, etc...

Does that make sense?

-Mike

[MarwaTawfik]

Thanks very much, Mike @SoilRotifer Yes it makes sense.