iTOL compatibility

[valentynbez]

Output trees from gtdb-tk are not compatible with iTOL, and iTOL is a very popular software for tree visualization. Standardization would help a lot.

[pchaumeil]

Hello,
A new functionality 'remove_labels' will be available in the next version of GTDB-Tk. This will remove all internal labels from a newick tree , making it compatible with Itol

[valentynbez]

I think bootstrap values would be also nice to have, maybe it would be possible to give a user control over the format.
iTOL accepts bootstrap variables as:
(A:0.1,(B:0.1,C:0.1)INT1:0.1[90])INT2:0.3[98])

• A, B, C : leaf names
• INT1, INT2 : internal node IDs
• 0.1, 0.3 : branch lengths
• 90,98 : bootstrap values

A regex could possibly be enough to change it.

[pchaumeil]

we have added a functionality to GTDB-Tk called 'convert_to_itol'. This command will convert a GTDB-Tk tree to an itol tree (with the format displayed above).
This option will be available in the next release of GTDB-Tk

[Biofarmer]

Hi, may I ask how to use 'convert_to_itol' to get itol tree, or the itol tree has been created in the output folder? Thanks

[pchaumeil]

Hi,
"convert_to_itol" command works as following:
gtdbtk convert_to_itol --input_tree INPUT_TREE --output_tree OUTPUT_TREE

where INPUT_TREE is the path to the tree in Newick format generated by GTDB-Tk and OUTPUT_TREE path to output the tree

[Biofarmer]

Hi, "convert_to_itol" command works as following: gtdbtk convert_to_itol --input_tree INPUT_TREE --output_tree OUTPUT_TREE

where INPUT_TREE is the path to the tree in Newick format generated by GTDB-Tk and OUTPUT_TREE path to output the tree

Hi, thanks for reply. May I ask which folder the Newick format tree generated by GTDB-Tk is located? There are 7 trees files (gtdbtk.bac120.classify.tree.1.tree,...,gtdbtk.bac120.classify.tree.7.tree) for bacteria and 1 tree file (gtdbtk.ar53.classify.tree) for archaea in 'classify' folder.
By the way, may I ask the difference between 'gtdbtk.bac120.user_msa.fasta.gz' and 'gtdbtk.bac120.msa.fasta.gz'? They are all protein sequence alignment, right?

Thanks

[pchaumeil]

There are 7 trees files (gtdbtk.bac120.classify.tree.1.tree,...,gtdbtk.bac120.classify.tree.7.tree) for bacteria and 1 tree file (gtdbtk.ar53.classify.tree) for archaea in 'classify' folder.
-you can pick whichever tree you want to convert. The classify command will produce 7 bacterial trees when using the split approach ( backbone + 6 class level tree) so you will need to pick your tree of interest out of those 7 . Those 7 tree are basically the GTDB reference tree split in smaller subtrees.
To have only one tree you can either run classify_wf with the --full_tree flag ( memory intensive) or the de_novo_wf command

By the way, may I ask the difference between 'gtdbtk.bac120.user_msa.fasta.gz' and 'gtdbtk.bac120.msa.fasta.gz'?
-'gtdbtk.bac120.user_msa.fasta.gz' is a MSA only with your genomes, 'gtdbtk.bac120.msa.fasta.gz' is a MSA with your genomes and the GTDB representatives ( larger file)

[Biofarmer]

There are 7 trees files (gtdbtk.bac120.classify.tree.1.tree,...,gtdbtk.bac120.classify.tree.7.tree) for bacteria and 1 tree file (gtdbtk.ar53.classify.tree) for archaea in 'classify' folder. -you can pick whichever tree you want to convert. The classify command will produce 7 bacterial trees when using the split approach ( backbone + 6 class level tree) so you will need to pick your tree of interest out of those 7 . Those 7 tree are basically the GTDB reference tree split in smaller subtrees. To have only one tree you can either run classify_wf with the --full_tree flag ( memory intensive) or the de_novo_wf command

By the way, may I ask the difference between 'gtdbtk.bac120.user_msa.fasta.gz' and 'gtdbtk.bac120.msa.fasta.gz'? -'gtdbtk.bac120.user_msa.fasta.gz' is a MSA only with your genomes, 'gtdbtk.bac120.msa.fasta.gz' is a MSA with your genomes and the GTDB representatives ( larger file)

thanks, if using ' --full_tree flag', bacteria and archaea will be also in one tree, or still each tree from them? Is gtdbtk.bac120.user_msa.fasta.gz protein sequence alignment?

[pchaumeil]

bacteria and archaea will always be on different trees ( their MSA, markers are completely different) and yes gtdbtk.bac120.user_msa.fasta.gz is a protein sequence alignement file

[Biofarmer]

There are 7 trees files (gtdbtk.bac120.classify.tree.1.tree,...,gtdbtk.bac120.classify.tree.7.tree) for bacteria and 1 tree file (gtdbtk.ar53.classify.tree) for archaea in 'classify' folder. -you can pick whichever tree you want to convert. The classify command will produce 7 bacterial trees when using the split approach ( backbone + 6 class level tree) so you will need to pick your tree of interest out of those 7 . Those 7 tree are basically the GTDB reference tree split in smaller subtrees. To have only one tree you can either run classify_wf with the --full_tree flag ( memory intensive) or the de_novo_wf command

By the way, may I ask the difference between 'gtdbtk.bac120.user_msa.fasta.gz' and 'gtdbtk.bac120.msa.fasta.gz'? -'gtdbtk.bac120.user_msa.fasta.gz' is a MSA only with your genomes, 'gtdbtk.bac120.msa.fasta.gz' is a MSA with your genomes and the GTDB representatives ( larger file)

Thanks. May I further ask if there are any differences between the 7 trees each other, and also the 7 trees comparing to the one tree if using --full_tree flag?

[pchaumeil]

To speed up the classification process, the full reference tree ( used by --full_tree) has been split in 6 smaller trees (class-level tree) , and each of them contains all representative genomes for a different set of class.
With the new split approach , each genome is now placed in a backbone tree ( tree containing only one genome per family) and, based on its placement and its RED value in this backbone tree, this genome will get a class rank classification.
Once the genomes get a class ( lets say class c__Bacteroidia) from the backbone tree , Tk will place them on a one of the 6 class level tree ( The class level tree containing c__Bacteroidia) to finalise the Taxonomy down to the species level

[Biofarmer]

To speed up the classification process, the full reference tree ( used by --full_tree) has been split in 6 smaller trees (class-level tree) , and each of them contains all representative genomes for a different set of class. With the new split approach , each genome is now placed in a backbone tree ( tree containing only one genome per family) and, based on its placement and its RED value in this backbone tree, this genome will get a class rank classification. Once the genomes get a class ( lets say class c__Bacteroidia) from the backbone tree , Tk will place them on a one of the 6 class level tree ( The class level tree containing c__Bacteroidia) to finalise the Taxonomy down to the species level

Many thanks.