Using Rstudio with mamba-installed R
__1. Download and unzip Rstudio Tarball
- Download tarball from : https://posit.co/download/rstudio-desktop/
- I downloaded the following tarball for my CentOS 7 server
- I downloaded the following tarball for my CentOS 7 server
- Unzip the tarball and change the owner and mode of chrome-sandbox
-
tar -zxf rstudio-2023.09.0-463-x86_64-fedora.tar.gz cd rstudio-2023.09.0+463 # the following 2 commands are from the error message that shows up after executing rstudio right after the tarball unzipping. sudo chown root:root chrome-sandbox sudo chmod 4755 chrome-sandbox ls -l chrome-sandbox #-rwsr-xr-x. 1 root root 53840 Sep 26 09:16 chrome-sandbox
-
__2. Download R using Conda or Mambaforge
- I prefer to use conda/mamba to install R and R packages for easy installation and compatibility.
- Mambaforge is better than plain conda nowadays because of the speed of package resolution and installation.
- How to install and setup mambaforge : https://github.com/conda-forge/miniforge
- create a conda environment and at the same time install R and R packages using mamba
-
mamba create -n cosmx -c conda-forge -c r -c bioconda r-base r-seurat r-devtools r-terra r-magick
-
__3. Environment Modules to easily set pathes
- In order for Rstudio to recognize the R installed using conda/mamba, we need to set PATH,LD_LIBRARY_PATH, and CPATH (I haven’t tested whether setting CPATH is essential or not)
- The Enviroment Modules package makes it easy to dynamically load and unload those path settings.
- Install Environment Modules : https://modules.readthedocs.io/en/stable/INSTALL.html
- install tcl and tcl-devel if not already before installation.
-
yum install tcl tcl-devel
-
- Download and install module files as described in the link : https://modules.readthedocs.io/en/stable/INSTALL.html
- To use modulefile you need to source the profile. For me, I put it in the ~/.bashrc. I also set my custom MODULEPATH where modulefiles are stored.
-
#~/.bashrc source /usr/local/Modules/init/profile.sh export MODULEPATH=${MODULEPATH}:/home/hd00ljy/modulefiles/tools
-
- make a modulefile
-
nano modulefiles/tools/cosmx/1
-
#%Module1.0 proc ModulesHelp { } { puts stderr " " puts stderr "\tThis module is for use of [module-info name]." puts stderr " " puts stderr "\tYou need module(s): " puts stderr "\t none " puts stderr "\tuse example: " puts stderr "\t $ module load [module-info name]" puts stderr " " } # install history ################################# #mamba create -n cosmx -c conda-forge -c r -c bioconda r-base r-seurat r-devtools r-terra r-magick ################################# #if { [ module-info mode load ] } { # module load gcc/7.4.0 #} module-whatis "loads the modules environment" # for Tcl script use only set name cosmx set version 4.3.1 set envdir /home/hd00ljy/mambaforge/envs set prefix $envdir/$name #if {![file exists $prefix]} { # puts stderr "\t[module-info name] Load Error: $prefix does not exist" # break # exit 1 #} prepend-path PATH $prefix/bin prepend-path LD_LIBRARY_PATH $prefix/lib/R/lib prepend-path CPATH $prefix/lib/R/include prepend-path LD_LIBRARY_PATH $prefix/lib prepend-path CPATH $prefix/include #if { [ module-info mode remove ] || [ module-info mode switch ] } { # module unload gcc/7.4.0 #}
-
- install tcl and tcl-devel if not already before installation.
- you can now load and unload conda-installed R(executables,libraries, and headers) using module command
-
module load cosmx/1 # load module module purge # unload all modules that have been loaded
-
__4. Execute Rstudio
- load module for conda/mamba-installed R
-
module load cosmx/1 # load module
-
- Launch Rstudio
-
cd rstudio-2023.09.0+463 ./rstudio
-