When developing R scripts for single-cell sequencing data analysis, utilizing Docker containers offers a streamlined and reproducible environment. This documentation outlines the process of developing R scripts within a Docker container, leveraging JupyterLab for efficient coding and visualization.

If the goal is to develop an R script named sc_singleR.prod.R, we create two files:

sc_seurat_opt.R   # https://github.com/CCRSF-IFX/SF_sc-smk-wl/blob/main/scripts/rna/sc_singleR_opt.R
sc_singleR.ipynb  # https://github.com/CCRSF-IFX/SF_sc-smk-wl/blob/main/scripts/rna/sc_singleR.ipynb

1. Record the script path:

script_path = getwd()                                # %exclude_jupyterlab%
script_path                                          # %exclude_jupyterlab%

1. Invoke sc_singleR_opt.R for command line options

Inside the notebook, we use the cell below to utilize sc_seurat_opt.R to read the command line options and save the information in an RDS file named opt.rds.

system(paste0('Rscript ', script_path,                # %exclude_jupyterlab%
        '/sc_singleR_opt.R --genome="hg38" --markerList="/Volumes/ccrsf-ifx/Software/scripts/bin/currentsnake/single_cell/gene_lists/human_gene_list.csv" --outdir="/Volumes/ccrsf-static/Analysis/xies4/github_repos/pipeline_dev_test/singleR" --rds="/Volumes/ccrsf-static/Analysis/xies4/github_repos/pipeline_dev_test/test_dir/seur_10x_cluster_object.rds"'),  # %exclude_jupyterlab%
       intern = T) # %exclude_jupyterlab%

1. Write the main code

After capturing the options, read them using readRDS() and proceed with the main code:

opt = readRDS("/Volumes/ccrsf-static/Analysis/xies4/github_repos/pipeline_dev_test/singleR/opt.rds")         # %exclude_jupyterlab%
# main code  goes here

1. Generate production script: sc_singleR.prod.R

Combine the necessary scripts and exclude irrelevant lines marked with exclude_jupyterlab to create the production script:

notebook_prefix = "sc_singleR"                                                         # %exclude_jupyterlab%
notebook_name = paste0(notebook_prefix, ".ipynb")                                     # %exclude_jupyterlab%
notebook_r = paste0(script_path, "/", paste0(notebook_prefix, ".r"))                  # %exclude_jupyterlab%
notebook_path = paste0(script_path, "/", notebook_name)                               # %exclude_jupyterlab%
opt_name = paste0(script_path, "/", sub(".ipynb", "_opt.R", notebook_name))           # %exclude_jupyterlab%
output = paste0(script_path, "/", sub(".ipynb", ".prod.R", notebook_name))            # %exclude_jupyterlab%
cmd1 = paste0("jupyter nbconvert --to script --output ",                              # %exclude_jupyterlab%
             notebook_prefix, ' ', notebook_path, "> /dev/null 2>&1 ")                # %exclude_jupyterlab%
cmd1                                                                                  # %exclude_jupyterlab%
system(cmd1, intern = TRUE)                                                            # %exclude_jupyterlab%
cmd2 = paste0('cat ', opt_name, ' ', notebook_r,                                      # %exclude_jupyterlab%
             ' |grep -v exclude_jupyterlab > ', output,  ' 2>&1')                     # %exclude_jupyterlab%
cmd2                                                                                  # %exclude_jupyterlab%
system(cmd2, intern = T)                                                              # %exclude_jupyterlab%
system(paste0("rm ", notebook_r))                                                     # %exclude_jupyterlab%