aforsythe/r-seurat

R-Studio / Seurat Docker Environment

---

This project is intended to allow users of R-Studio and Seurat to establish a reproducible environment for data analysis.

The general workflow is :

1. Run a docker container with R-Studio and Seurat
2. Create an analysis project structure and verison control repository
3. Create analysis scripts, functions, reports, etc. adding additional packages as necessary
4. Capture the R environment in a renv.lock file
5. Commit analysis to version control at appropriate time points

1. Run a docker container with R-Studio and Seurat

To run the docker container do the following in the host machine (e.g. Mac OS) terminal :

• Clone this repo (only need to do this the first time)
  • e.g. git clone https://github.com/aforsythe/r-seurat
• Run docker-compose for the version of Seurat you'd like to use.
  • e.g. cd dev && docker-compose up -d
• After the container has finished building, go to http://localhost:8787 in your browser
  • Building the container for the first time might take as long as 15 mins or so. Subsequent container launches will be much quicker, if not immediate.

2. Create an analysis project structure and verison control repository

To create an analysis project, type the following in the R console:

• project_template('my_analysis') where my_analysis is a name of your choosing for the project.

3. Create analysis scripts, functions, reports, etc. adding additional packages as necessary

All data in the container should be saved in the project directory (e.g. ~/r_data/my_analysis) The data will be available locally in ~/r_data/my_analysis. Save all files when stopping the container because container data is ephemeral.

• Create scripts in R and save to ./my_analysis/scripts
• Reusable functions should be saved to ./my_analysis/scripts/functions/
• Original data should be saved to ./my_analysis/data/original
• Cleaned data should be saved, using a script, to ./my_analysis/data/clean
• Any resulting data saved from a script should be saved to ./my_analysis/results/data
• Any resulting figures saved from a script should be saved to ./my_analysis/results/figures
• Rmarkdown reports should be saved to ./my_analysis/rmarkdown/reports
• Rmarkdown presentations should be saved to ./my_analysis/rmarkdown/presentations

4. Capture the R environment in a renv.lock file

When you've got an analysis that's working, make sure all scripts are saved then in the R console type:

• renv::snapshot()

5. Commit analysis to version control at appropriate time points

Commit the files to a git repo by typing the following in the terminal inside the container (e.g. using the terminal tab in RStudio)

• cd ~/r_data/my_analysis && git add .
• git commit -m 'my first commit'
  • If prompted, type: git config --global user.email "you@example.com" and git config --global user.name "Your Name" to setup git using your username and full name.
• Replace my first commit with an appropriate message describing what's changed in the code since the last commit.

When you are done with the project for a while, or want to start a new project, delete the Docker container by typing the following the in the Host terminal :

• docker stop $(docker ps -a -q)
• docker rm $(docker ps -a -q)

You can also use the stop and trash icons in the docker desktop gui.

To revisit a project:

• do step #1 (e.g. docker-compose up -d and visit https://localhost:8787)
• In RStudio go to File -> Open Project and navigate to the project you'd like to work with opening the .Rproj file (e.g. my_analysis.Rproj