Thanks and credit to Joern Toedling for this useful and clear how-to!
From The Bioconductor Digest, Vol 62, Issue 14:
You do not need to have root access to a machine to install your own
versions of R and your favourite packages there. This is how to do it:
1. downloaded the tar.gz of the development version of R from
ftp://ftp.stat.math.ethz.ch/Software/R/
2. uncompress it to a directory you have write access to, say ~/local/R
3. change into the uncompressed directory, ~/local/R/R-devel
4. run "./configure"
5. run "make"
Afterwards you can start R by executing ~/local/R/R-devel/bin/R;
to simplify that either add the bin directory to your path or create an alias for R
You do not need to run "make install" to work with R.
For packages,
1. create a directory in which you want the packages installed, e.g. ~/local/rpacks
2. create an evironment variable R_LIBS that holds the path to that directory, e.g. "setenv R_LIBS=${HOME}/local/rpacks" with that directory and a C-shell (use export with a Bash shell)
This environment variable tells R where to look first for installed packages and where to install packages when using "install.packages" or "biocLite".
R_LIBS is safe to use, since it only extends the path to look for packages and does not replace the default R library path.
I would recommend to add the alias for starting R and the R_LIBS
definition to your shell startup file (~/.cshrc or ~/.bashrc).
