This repo holds a Docker container with Rstudio server for scRNA-seq analysis.

Note

HPCs do not allow Docker containers, however, they are able to generate a singularity/apptainer image on the fly. Github only hosts Docker containers. Hence this workflow.

Running interactively

Important

The steps below are setup for running on the HPC UPPMAX Pelle.

The container are ran inside an interactive shell and forwarded from a login shell. When the container is running a password will be generated. Use it in step 6.

1. login to HPC
2. allocate a compute node

interactive -A "$PROJECT" -n 1 -t 1:00:00
salloc: Pending job allocation 4968700
salloc: job 4968700 queued and waiting for resources
salloc: job 4968700 has been allocated resources
salloc: Granted job allocation 4968700
salloc: Waiting for resource configuration
salloc: Nodes p115 are ready for job

3. start the conainer:

# On the HPC (from an interactive job or within a SLURM submission)
apptainer run \
  --cleanenv \
  --env RSTUDIO_PORT=8080 \
  --env RSTUDIO_ADDRESS=0.0.0.0 \
  docker://ghcr.io/karlssonlaboratory/rstudio-server:35b5461

4. open second terminal and ssh tunnel

#ssh -L <local_port>:<node_id>:<host_port>
ssh -L 8080:<compute-node-id>:8080 pelle

# for debug
#ssh -v -N -L 8080:<compute-node-id>:8080 pelle

5. Open a browser and run Rstudio

open http://localhost:8080

6. Use the generated password and your username