GTDB-Tk

A toolkit for assigning objective taxonomic classifications to prokaryote genomes

---

File naming

The file name must end with ".summary.tsv".

File format

tip

For more information on the GTDB-Tk output format, visit the GTDB-Tk documentation.

Results Summary (.tsv)

The file must include a header (i.e. the column names at the top). The column names can be anything, as long as the order is exactly the same. It must have the following columns, in that order:

Column name	Column obligatoriness	Data type	Data nullability
user_genome	Mandatory	String	Not nullable
classification	Mandatory	String	Not nullable
closest_genome_reference	Mandatory	String	Nullable
closest_genome_reference_radius	Mandatory	Float	Nullable
closest_genome_taxonomy	Mandatory (ignored)	N/A	N/A
closest_genome_ani	Mandatory	Float	Nullable
closest_genome_af	Mandatory	Float	Nullable
closest_placement_reference	Mandatory	String	Nullable
closest_placement_radius	Mandatory	Float	Nullable
closest_placement_taxonomy	Mandatory (ignored)	N/A	N/A
closest_placement_ani	Mandatory	Float	Nullable
closest_placement_af	Mandatory	Float	Nullable
pplacer_taxonomy	Mandatory (ignored)	N/A	N/A
classification_method	Mandatory	String	Not nullable
note	Mandatory	String	Nullable
other_related_references	Mandatory (ignored)	N/A	N/A
msa_percent	Mandatory (ignored)	N/A	N/A
translation_table	Mandatory (ignored)	N/A	N/A
red_value	Mandatory	Float	Nullable
warnings	Mandatory	String	Nullable

info

Why are there mandatory columns that are ignored?

That has to do with the way the GTDB-Tk file parser is written. When the file is read, it must comply with a pre-defined schema (column order and types), even though some of these columns end up being dropped later.

Mapping to database

GTDBTkTsvFile

Original data	GTDBTkTsvFile field	Notes
GTDB-Tk file path	path

GTDBTkTsvEntry

Original data	GTDBTkTsvEntry field
user_genome	genome_key 1
classification	domain, phylum, klass, order, family, genus, species 2
closest_genome_reference or closest_placement_reference column	reference 3
closest_genome_reference_radius or closest_placement_radius column	radius 3
closest_genome_ani or closest_placement_ani column	ani 3
closest_genome_af or closest_placement_af column	af 3
classification_method	classification_method
note	note
red_value	red_value
warnings	warnings

Footnotes

1. The MAG name in the GTDB-Tk file name is used to query the primary key of the corresponding genome in the database. ↩

2. The classification column is broken down into multiple fields for better readability. ↩

3. The closes_placement_* columns are only filled when the classification method used by GTDB-Tk is ANI screen. Otherwise, the closest_genome_* columns are filled. With that in mind, parsomics includes only the relevant metrics to each classification method. ↩ ↩² ↩³ ↩⁴