Building a database
Building a database in seed-Kraken requires providing a spaced seed
as an argument: --seed SEED
,
A seed sequence is then stored in a file inside the database directory: seed-kraken-db/spaced_seed_sequence
.
Once built the database will work only with this seed.
Download options: --download-library
, --download-taxonomy
do not need the seed argument.
You will find a more detailed description of the building process in README of the original Kraken.
Bash script
After downloading taxonomy, and downloading/collecting the library, this example database building bash script can be run.
#!/bin/bash
set -u # Protect against uninitialized vars.
set -e # Stop on error
set -o pipefail # Stop on failures in non-final pipeline commands
KRAKEN_DB_NAME="$1
SEED="######-##-#-#-##-###-#-##---###--#######"
THREADS=20
MINIMIZER_LEN="12"
JELLYFISH_HASH_SIZE="5000M"
check_for_jellyfish.sh
kraken-build --db $KRAKEN_DB_NAME --build --threads $THREADS -seed $SEED \
--minimizer-len $MINIMIZER_LEN --jellyfish-hash-size $JELLYFISH_HASH_SIZE