Installing R Packages on Fox

R is a programming environment for performing statistical operations.

To find out more, visit the R website at: https://www.r-project.org

Selecting the correct module to load

To see available versions when logged into Fox issue command

module spider R

To use R type

module load R/<version>

How to install packages as a user

There might be some packages missing in the R module we have installed or may be you need a different version than what we have. In that case you could install the packages yourself. For example, following is the procedure to install the package called XYZ by the user ec-user1 on Fox. Please remember to use your username instead of ec-user1

• Login to Fox
• Load the module
• Create a directory to place the libraries
• Start R
• Set the library path
• Install the package

        [ec-user1@login-4 ~]$ module purge
        [ec-user1@login-4 ~]$ module load R/4.0.0-foss-2020a
        [ec-user1@login-4 ~]$ mkdir /cluster/home/ec-user1/R
        [ec-user1@login-4 ~]$ R

Use the R prompt to install the package

      #Set the location for the packages to be installed
      > .libPaths(c("/cluster/home/ec-user1/R",.libPaths()))
      #install the package
      >  install.packages("XYX", repo="cran.uib.no")
      #Check if the package can be  loaded
      > library(XYZ)

How to use an installed package After installing, everytime the packages needed to be accessed The .libPaths("/cluster/home/ec-user1/R") setting should be done. When submitting R Script as a job, the .libPath("/cluster/home/ec-user1/R") should be specified before calling the package.

Rscript example in a job

In order to submit as a job, we need to include all R processing steps in an Rscript

Example Rscript

[ec-user1@login-4 ~]$ cat test.rscript 

#!/usr/bin/env Rscript

.libPaths(c("/cluster/home/ec-user1/R",.libPaths()))
library("ggplot2")
out_put = paste("Processing started on", Sys.info()["nodename"])
paste(out_put)

mean_plus_this <- function(data, addthis = 0) {
  new_data <- mean(data) + addthis
  return(new_data)
}

test_data <- c(4, 40, 9, 1)
mean_plus_this(test_data, 3)

Example job script to call the above Rscript

#!/bin/bash

#SBATCH --account=ec12
#SBATCH --job-name=R_script_test
#SBATCH --mem=8G
#SBATCH --ntasks=1
#SBATCH --time=15:00

set -o errexit # Make bash exit on any error
set -o nounset # Treat unset variables as errors

module purge
module load R/4.0.0-foss-2020a

Rscript test.rscript > output.Rout 2> error.Rout

Bioconductor

We have also made bioconductor as a module with the base bioconductor packages. Not all packages found in biocondctor repository are pre-installed. This is becasue compatibility issues, i.e. not all packages can be installed with the same set of dependancies and to make it easier for the user to select the exact version combinations for some packages.

Selecting the correct module load

To see available versions when logged into Fox issue command

module spider bioconductor

If you do not have specific requirements please use the latest, as of now it is "R-bundle-Bioconductor/3.12-foss-2020b-R-4.0.3"

When you load the Bioconductor module a compatible R module and all supporting module that are needed to build libraries will be loaded.

Here is an example showing how to install a bioconductor package

        [ec-user1@login-4 ~]$ module purge
        [ec-user1@login-4 ~]$ module load R-bundle-Bioconductor/3.12-foss-2020b-R-4.0.3
        [ec-user1@login-4 ~]$ mkdir /cluster/home/ec-user1/R
        [ec-user1@login-4 ~]$ R

• Use the R prompt to install the package

      #Set the location for the packages to be installed
      > .libPaths(c("/cluster/home/ec-user1/R",.libPaths()))
      #install the package
      >  BiocManager::install("XYZ", repo="cran.uib.no")
      #Check if the package can be  loaded
      > library(XYZ)

License Information

R is available under several open-source licenses. For more information, visit https://www.r-project.org/Licenses/

It is the user's responsibility to make sure they adhere to the license agreements.

^{CC Attribution: This page is maintained by the University of Oslo IT FFU-BT group. It has either been modified from, or is a derivative of, "Installing R packages" by NRIS under CC-BY-4.0.}