Get a reference phylogenetic tree of known taxa from GTDB

Question

[Gio]

Hello,

I have a set of genomes I downloaded from NCBI. I would like to make a reference phylogenetic tree where only they appear.

Instead of aligning them or using mash distance to make my own tree, is there a way I can simply provide the genomes or taxa to GTDB and get a tree back from it?

Answer 1

[Mensur Dlakic]

To the best of my knowledge, there is no way to do this using GTDB website. However, it can be done locally using the GTDB-toolkit. You select a group of genomic sequences and run it through the program. It will do gene predictions, single-copy marker identification, and a taxonomic assignment that will include a global GTDB tree with your organisms (too big for most applications) plus a tree with just your organisms.

https://github.com/Ecogenomics/GTDBTk

Answer 2

[fredjaya]

If I understand correctly, given a list of the NCBI genome IDs, you can prune a .tree (e.g. the GTDB tree) with Biopython. Note that the NCBI IDs must match the taxa labels in the tree.

Example script here.

[Mensur Dlakic]

Nicely done. Yet pruning the tree will not suffice for the user-specific entries that may not already be in the tree.

[fredjaya]

Good point. I think it's possible to "place" tips via IQ-TREE2's contrained tree search option.

iqtree -s user_sequences.fa -g pruned_gtdb.tree

Here the starting tree is fixed to the pruned tree, and will infer the whole tree to include the "new" user-specific sequences.

[Andrzej Zielezinski]

I think it is highly likely that most (or all) user-specified NCBI genomes can be found in GTDB. The reference GTDB tree encompasses approximately 80,000 bacterial species, each represented by a single genome (representative genome for species). However, the complete GTDB database comprises around 400,000 genomes, including both representative and non-representative genomes. To access the full genome list, including NCBI identifiers, the user can download the metadata file from GTDB at https://data.gtdb.ecogenomic.org/releases/latest/. This file provides details for all 400,000 genomes present in the GTDB database.