CADD docker image on Singularity

[e-271]

Hello,

I am trying to run the Docker image provided for CADD1.7 using Singularity on a HPC:

https://hub.docker.com/layers/visze/cadd-scripts-v1_7/0.1.1/images/sha256-5774cde9ab918c0726ab0cf443b25789b0d8bea2deb1173772ba6e2707528a2c

However it doesn't seem to have CADD or the required dependencies installed when I run the image.

I am running the following:

singularity pull docker://visze/cadd-scripts-v1_7:0.1.1
singularity shell cadd-scripts-v1_7_0.1.1.sif
git clone https://github.com/kircherlab/CADD-scripts
./CADD-scripts/CADD.sh -a CADD-scripts/test/input.vcf

And I get this error, it seems Snakemake is not installed (and I also don't find an executable command for CADD without cloning the repository):

Singularity> ./CADD-scripts/CADD.sh -a CADD-scripts/test/input.vcf
CADD-v1.7 (c) University of Washington, Hudson-Alpha Institute for Biotechnology and Berlin Institute of Health at Charite - Universitatsmedizin Berlin 2013-2024. All rights reserved.
Running snakemake pipeline:
snakemake /tmp/tmp.fcGgOPNCkn/input.tsv.gz --sdm conda apptainer --apptainer-prefix /home/erobb/CADD-scripts/CADD-scripts/envs/apptainer --singularity-args "--bind /tmp/tmp.fcGgOPNCkn " --conda-prefix /home/erobb/CADD-scripts/CADD-scripts/envs/conda --cores 1 --configfile /home/erobb/CADD-scripts/CADD-scripts/config/config_GRCh38_v1.7.yml --snakefile /home/erobb/CADD-scripts/CADD-scripts/Snakefile -q
./CADD-scripts/CADD.sh: line 155: snakemake: command not found

Is Singularity not supported for the Docker image? Or do I need to do some additional setup?
Thank you!

[visze]

Hi,

this is not a container that includes snakemake and CADD-scripts. This container includes only the conda environments needed for the different snakemake rules (see here).

So you can run snakemake without a local conda installation. This command snakemake --sdm apptainer conda will download the image and runs for each rule the script within the image actvating the conda environment. If you run it via CADD.sh you need to run it without the option -m. But snakemake has to be installed.

So please follow the readme in the readme in this repository to get CADD-scripts running local in your environment.

Best,
Max

[e-271]

Got it, thank you!

If possible I think it'd be helpful to release a container that has Snakemake installed and CADD available as an executable, for streamlining in larger pipelines.

[visze]

Maybe it helps. We have something like this:

visze/cadd-scripts-v1_7

For cadd v1.7. but right now no documentation and and therefore no support yet. Some temp dirs and the annotation as well pre-scored files have to be correctly bind.

[e-271]

Okay thank you. Is that container available for download anywhere? I've been using visze/cadd-scripts-v1_7:0.1.1, and tried docker pull visze/cadd-scripts-v1_7 but that seems to point me to 0.1.1.

Thank you!

[visze]

Yes it is on docker hub.

[e-271]

It seems to be the same have been using. Please let me know if there is another updated image I should use, otherwise I'll set it up custom. Thank you.

[visze]

Oh yes! My fault. It is visze/cadd-scripts:1.7 on docker hub for v1.7 and visze/cadd-scripts:1.6 for v1.6. all other versions are also available.