The annotation results of gtdbtk are inconsistent with the Phylogenetic position.

[wanxn518]

Hello, GTDB team

I use GTDBtk version 2.3.2 to classify my MAGs, so I know that these MAGs belong to a certain phylum and a certain class. However, in the phylogenetic tree, they are inferred to be other classes in the same phylum. Is this because the MAGs was incorrectly classified when it run classify_wf? Can you give me some suggestions? Thank you sincerely in advance!

Best, xnw

[donovan-h-parks]

Hi xnw,

Can you provide us with some more details? Which phylogenetic tree are you referring to? A tree produced by the GTDB-Tk de novo workflow? A tree you inferred?

The GTDB-TK classify_wf determines taxonomic assignments by placing your genomes into a fixed reference tree. The GTDB-Tk de novo workflow infers a new tree and doesn't directly provide any taxonomic classifications. It is possible that the results of these two inference methods will differ in some cases.

Thanks, Donovan

[wanxn518]

Hi Donovan,

Thank you very much for your reply. I've identified a number of classifications of MAGs through GTDBtk. gtdbtk classify_wf --genome_dir MAGs/ --out_dir GTDB_Result -x fna --prefix KON --cpus 24 --skip_ani_screen The classification results of these MAGs belong to p__Actinomycetota,c__Geothermincolia in bac120.summary.tsv.

I used these MAGs to make phylogenetic trees run gtdbtk de_novo_wf However, in the phylogenetic tree, the phylogenetic positions of individual MAGs are at p__Actinomycetota & c__Actinomycetia，and is also inferred in the de_novo_wf/infer/gtdbtk.bac120.decorated.tree-taxonomy as d__Bacteria; p__Actinomycetota; c__Actinomycetia; o__; f__; g__; s__

These are a few MAGs whose classification results from gtdbtk classify_wf don't agree with the phylogenetic position, and I'm not sure what their classification results actually are? Looking forward to your help and sincerely appreciate it in advance!

Best, xnw

[donovan-h-parks]

Hi xnw,

I assume your MAGs are the red entries marked with "???" - yes? It appears the challenge here is that in the de novo tree Actinomycetia is no longer a monophyletic group. Your MAGs are sister to Geotherminicolia and could reasonably be consider the most basal lineage of this class if it weren't for the small subset of Actinomycetia that have moved and are causing Actinomycetia to be polyphyletic. This is, unfortunately, the challenge of doing taxonomic classification with a de novo tree. As it stands, there is no clear classification for your MAGs. They could be considered a basal lineage of Geothermincolia or a new class within the Actinomycetota. Additional trees using different marker genes, evolutionary models, and inference methods are required to resolve which of these scenarios is best supported.

Cheers, Donovan

[phugenholtz]

Could you send me a high resolution image of your tree please xnw (so I can read the tip names). I suspect this is a know problematic region and I may have a solution.Thx, PhilSent from my iPhoneOn 22 Nov 2023, at 00:48, Donovan H. Parks @.***> wrote: Hi xnw, I assume your MAGs are the red entries marked with "???" - yes? It appears the challenge here is that in the de novo tree Actinomycetia is no longer a monophyletic group. Your MAGs are sister to Geotherminicolia and could reasonably be consider the most basal lineage of this class if it weren't for the small subset of Actinomycetia that have moved and are causing Actinomycetia to be polyphyletic. This is, unfortunately, the challenge of doing taxonomic classification with a de novo tree. As it stands, there is no clear classification for your MAGs. They could be considered a basal lineage of Geothermincolia or a new class within the Actinomycetota. Additional trees using different marker genes, evolutionary models, and inference methods are required to resolve which of these scenarios is best supported. Cheers, DOnovan

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: @.***>