picard-docker-singularity

Build Docker container for Picard Tools and (optionally) convert to Apptainer/Singularity.

Picard is a set of command line tools for manipulating high-throughput sequencing (HTS) data and formats such as SAM/BAM/CRAM and VCF.

Requirements:

Java 17
R

Build docker container:

1. For Picard installation instructions:

https://broadinstitute.github.io/picard/

2. Build the Docker Image

To build image from the command line:

# Assumes current working directory is the top-level picard-docker-singularity directory
docker build -t picard:3.0.0 . # tag should match software version

• Can do this on Google shell

To test this tool from the command line:

docker run --rm -it picard:3.0.0 java -jar picard.jar -h
docker run --rm -it picard:3.0.0 java -jar picard.jar MarkDuplicates --version

Optional: Conversion of Docker image to Singularity

3. Build a Docker image to run Singularity

(skip if this image is already on your system)
https://github.com/mattgalbraith/singularity-docker

4. Save Docker image as tar and convert to sif (using singularity run from Docker container)

docker images
docker save <Image_ID> -o picard3.0.0-docker.tar && gzip picard3.0.0-docker.tar # = IMAGE_ID of Picard image
docker run -v "$PWD":/data --rm -it singularity:1.1.5 bash -c "singularity build /data/picard3.0.0.sif docker-archive:///data/picard3.0.0-docker.tar.gz"

NB: On Apple M1/M2 machines ensure Singularity image is built with x86_64 architecture or sif may get built with arm64

Next, transfer the picard3.0.0.sif file to the system on which you want to run Picard from the Singularity container

5. Test singularity container on (HPC) system with Singularity/Apptainer available

# set up path to the Singularity container
PICARD_SIF=path/to/picard3.0.0.sif

# Test that Picard can run from Singularity container
singularity run $PICARD_SIF java -jar /picard.jar -h # depending on system/version, singularity may be called apptainer