Kraken

Technically its Kraken2, but I'm dropping the 2 for this.

kraken2 - bracken

Kraken's "standard" library contains archaea, bacteria, viral, human and the UniVec Core. If human is present, it usually doesn't find others so I created "abv" which is the same as "standard" but without human. Kraken's building process dust mask's the sequences, but does not repeat mask it. Having run it and found a lot of BeAnn, I think that I should run repeat masker on the sequences and rebuild.

kraken2-build --db standard-20200615 --download-taxonomy --skip-maps
kraken2-build --db standard-20200615 --download-library archaea
kraken2-build --db standard-20200615 --download-library bacteria
kraken2-build --db standard-20200615 --download-library viral
kraken2-build --db standard-20200615 --download-library UniVec_Core

#	Repeat Masker on the 4 library.fna files

kraken2-build --db standard-20200615 --build --threads 32


#	Prepare Bracken reference for each anticipated read length
#	Run multiple, but not at the same time.
#	They both create the file `database.kraken` if it doesn't exist.

bracken-build -d abv -t 32 -l 150

#	Then

bracken-build -d abv -t 32 -l 100



#	Then align with kraken with the report option and analyze the report file with bracken.

kraken2 --db ${KRAKEN_DB} --threads ${threads} --output ${f} --paired --use-names ${r1} ${r2}

bracken -d ${KRAKEN_DB} -i ${SAMPLE}.kreport -o ${SAMPLE}.bracken -r ${READ_LEN} -l ${LEVEL} -t ${THRESHOLD}