Kraken(2) database based on the GTDB project
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data added build from source instructions Dec 10, 2018
scripts fixed minor bug; added unit test Dec 10, 2018
tests heavy comments Dec 10, 2018
README.md error 429 Dec 10, 2018

README.md

GTDB_Kraken

Kraken(2) database based on the GTDB project

This page will contain links, notes and scripts for kraken databases we produced based on the GTDB database (http://gtdb.ecogenomic.org/). The kraken database contains 1 genome per taxon recognized in the GTDB database. Given that the the taxonomy of the GTDB database is based on genomic data, we expect this database to result in better performance for taxonomy-based classifier algorithms.

Version 1

Notes:

  • Based on GTDB Release 03-RS86 (19th August 2018)
  • Not the complete GTDB database, taxa without a genus designation have been omitted (10,994 out of 11042 taxa in this version)
  • Reference sequences have been selected ad hoc, without regard for the quality of the sequence. This will be dealt with in future versions

Kraken2 db (65 Gb, tarball 48 Gb):

gtdbk2_bacterial_v1.tar.gz md5sum: eda7855cb38a14b4222381ac5b27fe4b

https://drive.google.com/file/d/18E0W_ezNLAhxxZjjelQYwoLA_0oBm4Lo/view?usp=sharing

Building from source

To download all fasta files and compile from source, you can run the provided script with the input file from the GTDB project. Next, use the --add-to-library and --build functions in Kraken to format the database. Example commands are below.

Requirements

Example commands for building from source

# Creates the directory library/gtdb and adds fasta files.
# Also creates a taxonomy folder, compatible with Kraken.
perl scripts/gtdbToTaxonomy.pl --infile data/gtdb.2018-12-10.tsv
# Format the database.
db=GTDB_Kraken
for i in library/gtdb/*.fna; do 
  kraken-build --add-to-library $i --db $db
done
mv -v taxonomy $db
kraken-build --build --db $db --threads 8

Cleanup (optional)

You can optionally remove all intermediate fasta files and also run the clean utility in Kraken.

rm -rvf library
kraken-build --clean $db

Known issues

  • Error 429 too many requests. NCBI is receiving too many general requests, but you can carve out a special place for yourself by getting an NCBI API key. Log into your NCBI profile and copy your key. Then, add it to your environment like so: export NCBI_API_KEY=1fe2...

Authors

  • Henk den Bakker @hcdenbakker
  • Lee Katz @lskatz