Skip to content

Latest commit

 

History

History
61 lines (34 loc) · 2.33 KB

rstudio_sc_docker.md

File metadata and controls

61 lines (34 loc) · 2.33 KB
Raw

Docker image with rstudio for single cell analysis

Description

Docker images for single cell analysis.

All docker images contain an rstudio installation with some helpful packages for singlecell analysis. It also includes a conda environment to deal with necessary python packages (like umap-learn).

Docker Rstudio images are obtained from rocker/rstudio.

R version and Bioconductor

The R and Bioconductor versions are specified in the image name (along with the OS version):

Example: singlecell-base:R.4.0.3-BioC.3.11-ubuntu_20.04

Use

docker run -d -p 8787:8787 --name <container_name> -e USER='rstudio' -e PASSWORD='rstudioSC' -e ROOT=TRUE -v <host_folder>:/home/rstudio/projects vbarrerab/<docker_image>)

-e DISABLE_AUTH=true option can be added to avoid Rstudio login prompt. Only use on local machine.

This instruction will download and launch a container using the singlecell image. Once launch, it can be access through a web browser with the URL 8787:8787 or localhost:8787.

Important parameters

  • -v option is mounting a folder from the host in the container. This allows for data transfer between the host and the container. This can only be done when creating the container!

  • --name assigns a name to the container. Helpful to keep thins tidy.

  • -e ROOT=TRUE options provides root access, in case more tweaking inside the container is necessary.

  • -p 8787: Change the local port to access the container. This can only be done when creating the container!

  • FYI: The working directory will be set as /home/rstudio, not /home/rstudio/projects as default behavior.

Resources

The dockerfile and other configuration files can be found on:

https://github.com/vbarrera/docker_configuration

The docker images:

vbarrerab/singlecell-base

Available images:

  • R.4.0.2-BioC.3.11-ubuntu_20.04
  • R.4.0.3-BioC.3.11-ubuntu_20.04

Important:

Docker changed its policies to only keep images that have been modified in the last 6 months. This means that previous images will eventually disappear. For previous versions. Check with availability with @vbarrera.

Bibliography

Inspired by:

https://www.r-bloggers.com/running-your-r-script-in-docker/

Other resources

Using Singularity Containers on the Cluster: https://docs.rc.fas.harvard.edu/kb/singularity-on-the-cluster/