Running VEP on CREATE

Variant Effect Predictor (VEP) is a Ensembl provided tool and can be ran as a containerized application on CREATE HPC through Singularity.

Setting up your container

Take note

Singularity is only accessible from the compute nodes, to use VEP on the HPC, please either request an interactive session or submit the application as a batch job.

k1234567@erc-hpc-login:~$ srun -p cpu --time 60 --pty /bin/bash -l
k1234567@erc-hpc-comp1:~$ singularity --help

A Docker image is available for VEP, and this can be converted and used through Singularity:

singularity pull docker://ensemblorg/ensembl-vep

This will create a new .sif container called: ensembl-vep_latest.sif and can be called with instructions similar to the following:

singularity exec ensembl-vep_latest.sif /opt/vep/src/ensembl-vep/vep --help

Downloading and installing cache data

The following example, made applicable to CREATE HPC, is based on the corresponding FASTA for human GRCh38 outlined in the official Ensembl VEP documentation here.

Included in the image is the installer script that can be used to download and setup your required cache data, and before you action the download you must first setup a non-default HPC cache directory to avoid default storage issues, this requires the binding of your HPC directory to a Singularity bind variable so that HPC path is within the same context as your container and therefore in scope of your VEP application:

singularity exec -B /scratch/users/k1234567/vep-data:/opt/vep/.vep ensembl-vep_latest.sif \
    perl /opt/vep/src/ensembl-vep/INSTALL.pl \
    -c /opt/vep/.vep \
    -a cf \
    -s homo_sapiens \
    -y GRCh38

The required data will be installed to the container mounted directory: /opt/vep/.vep, bound to example HPC path: /scratch/users/k1234567/vep-data.

If you were to instead manually download the cache data, with reference to the Ensembl annotation sources, with available linux tools, such as: wget or curl, you will just need to make sure you direct and bind your Singularity container to the relevant path where your manual download is located. To test your VEP container processes, the following examples are available:

singularity exec -B /scratch/users/k1234567/vep-data:/opt/vep/.vep ensembl-vep_latest.sif \
    vep --dir /opt/vep/.vep \
    -i /opt/vep/src/ensembl-vep/examples/homo_sapiens_GRCh38.vcf \
    --cache