makeblastdb4cloud

This document describes how to create custom blast databases using Docker container.

Sample command line

docker run \
    -v /home/raytseli/google-key.json:/etc/google-key.json \
    -v /home/raytseli/t.prot:/tmp/t.prot \
    ncbi/makeblastdb4cloud \
    make_custom_blastdb.sh \
    prot \
    /tmp/t.prot \
    gs://yan-customdb-test/d5 \
    aprot5 \
    "protein database 5"

Explanation

docker run -- invocation of the Docker container

-v /home/raytseli/google-key.json:/etc/google-key.json -- This parameter maps external to docker file, /home/raytseli/google-key.json, to the file /etc/google-key.json accessible in the docker container. The file must contain authentication credentials which allow the read/write access to the GCS bucket used to place custom blast databases. Note that the process running in the docker expects that credentials file will be found at /etc/google-key.json, so the external to docker file must be mapped to this specific location. If this parameter is not passed to docker, or the credentials in the file are invalid for read/write access to GCS bucket, the process will fail unless the GCS bucket in question is publicly readable and writable (see Google documentaion on setting GCS bucket permissions).

-v /home/raytseli/t.prot:/tmp/t.prot -- This parameter maps external to docker sequence file, /home/raytseli/t.prot, to the file /tmp/t.prot accessible in the docker container. The file must contain either sequence ids, one per line, or sequences in FASTA format.

yraytselis/make-custom-db:v5 -- This parameter is used to retrieve from docker hub the docker image used for creating docker container.

make_custom_blastdb.sh -- This parameter invokes the main driver script within docker container used for custom database creation.

prot -- sequence type, either prot or nucl.

/tmp/t.prot -- sequence file as specified by the file mapping above.

gs://yan-customdb-test/d5 -- GCS bucket and directory where custom daatabase will be placed. If only GCS bucket is specified, then the script will create a new directory in this bucket with the current timestamp as the name.

aprot5 -- name of the custom database.

"protein database 5" -- optional database title. If title is not provided, it will be the same as the database name.