16s circular phylogenetic tree

Dear qiime 2 team and communities,
Thank you for maintaining and updating the forum meticulously. As a mammalian evolutionary geneticist with not yet enough experience in 16s metagenomics I have a very basic question:

I see is that in our analysis and many other microbial analysis the circular phylogenetic tree group a single bacterial group in two or several different sections.

e.g, In the attached picture blue bacteria (proteobacteria) are found in different places in the
circular tree. It means on the surface that some blue bacteria (proteobacteria) are actually more closely related to the pink bacteria (acidobacteria) than to other proteobacteria ? Is this pattern due to paraphyletic origin of some species or maybe contentious inconsistency between traditional and sequence based taxonomy in bacteria? I appreciate your opinion. I am sure there is something interesting here that I don’t know.

Thanks in advance!

Picture/citation can be found alternatively at:
https://www.researchgate.net/post/metagenomic_16s_phylogenetic_trees[Circular-maximum-likelihood-phylogenetic-tree-based-on-OTU-representative-sequences-of|560x500]

(upload://fw4XNLXvdu4CkrXBMwNUm16Z9Ld.jpeg)

HI @arsalane1978,

I haven't been able to see your tree (it just redirects to a research gate topic page, unfortunately), so Im making my best guess.

I'd argue so. (Although as far as I can tell, this is a problem across evolutionary biology. I had a similar discussion with a plant ecologist last month.) I think it may be more complicated in bacteria, but I havent worked in macroecology, so this is a best guess.

There's very little bacteria that can be isolated with traditional methods, and otherwise an anerobic chamber and special equipment is required. So, its a lot harder for us to say, "This sample is a rhino :rhinoceros:, This sample is a hippo :hippopotamus:, This sample is an elephant :elephant:", We're dealing with an ensamble.

In most cases, we're assigning taxonomy with a naive baysian classifier, which ends up being only as good as the reference we're using. Since references are often built off ensamble because culturing is difficult, it becomes a circular problem. Add in the problem that a lot of de novo approaches relied on Fast Tree, which built a phylogenetic tree off short MiSeq reads, and there are potential questions about the natures of the tree as well.

I don't say this to discourage you, more in hopes that it lays out some of the challenges of the field. At this point, we're young, we're noising, and w'ere probably wrong on multiple levels. But, I guess Id argue not to give up, and just be cautious with inferences based on taxonomy alone.

I hope (:crossed_fingers:) this is a reasonable answer. Id love to learn more, though!

Best,
Justine

2 Likes