db-kraken2-custom

Container

from docker img: quay.io/biocontainers/kraken2:2.1.3--pl5321hdcf5f25_1 (!! too big img 333MB !!)

ToDo: Create a new one dedicated to "simple" index building from 'quay.io/nexomis/kraken2:2.1.3' and including important missing tools (rsync, tools for remove low complexity, ... but no need to include all recquired tools (ex: peptide sequence inclusion tools))

CURRENT STATUS : under_dev | under_test | in_prod

(not all features and feature combinations have yet been fully tested)

Description

Taxonomy

Arguments: -x
If you have already downloaded the kraken2 taxonomy database, you can set its path with this argument. This will lead to a (complete*) copy of the database in <out_db_dir> and avoid downloading (--download-taxonomy).
For information, the complete database after --download-taxonomy (downloaded, decompressed and processed) currently weighs 40G.0 **probably improved by copying only the information relative to interest taxon/accesion_id or by using a symbolic link instead of complete copy(ln -s instead of cp -R)*.

Nucleotide Sequences

Nucleotide sequence incorporation into the newly created database can be achieved in 3 ways, which can be used alone or in combination:

option 1:

Arguments: i and -n
Import genomic and transcriptomic sequences from 'ftp.ncbi.nlm.nih.gov/genomes/all/':

• *_genomic.fna.gz: top-level (exhaustive without unless redundancy) but repeatitive sequences are soft masked (if need to unmask, download manually and perform awk '{if(/^[^>]/)$0=toupper($0);print $0}' genomic.fna > genomic.unmasked.fna and submit to this script via option 3).
• *_rna_from_genomic.fna.gz: no need to concatenate with _rna.fna.gz as it must be included in rna_from_genomic.fna.gz (?). In addition '_rna.fna.gz' seems to be specific to refSeq ftp (not in genbank ftp))

option 2:

Argument: -l
Kraken2 import of predefined library. Current available option: 'archaea', 'bacteria', 'plasmid', 'viral', 'human', 'fungi', 'plant', 'protozoa', 'nr', 'nt', 'UniVec', 'UniVec_Core'.
Note: Only the genome seems to be incorporated (at least this is the case for 'human').

option 3:

No specific arguments, but the -f option must be used (with the risk of data loss !!).
The fasta to be integrated must be placed in the temporary folder <out_db_dir>/library/add_custom_tmp/.
They must be in the right format, especially for the ability to attribute them taxonomically by means of a valid accession id (NCBI) or the addition of a taxon id in the name of each sequence. See kraken2 manual:

Sequences not downloaded from NCBI may need their taxonomy information assigned explicitly. This can be done using the string kraken:taxid|XXX in the sequence ID, with XXX replaced by the desired taxon ID. For example, to put a known adapter sequence in taxon 32630 ("synthetic construct"), you could use the following:

>sequence16|kraken:taxid|32630  Adapter sequence
CAAGCAGAAGACGGCATACGAGATCTTCGAGTGACTGGAGTTCCTTGGCACCCGAGAATTCCA

Exemple usage:

Only Human:

option 1 (genome + transcriptome): ./build_custom_db.sh -i GCF_000001405.40 -n GRCh38.p14 -o kraken2DB_GRCh38 -t 8

option 2 (genome): ./build_custom_db.sh -l human -o kraken2DB_human -t 8

Human + bacteria and plasmid kraken library:

option 1 (human) + option 2 (bacteria and plasmid) ./build_custom_db.sh -i GCF_000001405.40 -n GRCh38.p14 -o kraken2DB_GRCh38_bacteria -t 8 -l bacteria,plasmid

option 2 (genome): ./build_custom_db.sh -l human,bacteria,plasmid -o kraken2DB_human_bacteria -t 8

Potential interest kraken2-build option :

Managable by argument in section Kraken2-Build Arguments

• --download-library/--download-taxonomy: maybe is better with --use-ftp (instead of rsync)
• --add-to-library: --no-masking (e.g: for classic virus this simplify dowstream assembly whithout risk to exclude viral reads at kraken host exclussion step and probably accelerade this step of kraken2-index).
• --build: --fast-build and --skip-maps

Potential improvement:

• create specific docker img ?
• improve taxonomy directorie copy : symbolic ling OR extraction of only interest features
• automatise option3:
  • cp input fasta path on good directories
  • add correspond taxid to all sequence name

With docker

docker run -it -u $UID:$GID -v $PWD:$PWD -w $PWD quay.io/biocontainers/kraken2:2.1.3--pl5321hdcf5f25_1 bash build_custom_db.sh -o k2_bact_arch -t 8 -l archaea,bacteria -f