Hi,
I’m using biomod2 package v4.2.6.2 in R on Baobab, and my script works except for GAM models. I need mgcv package, but with R 4.3.2, mgcv won’t install because R is too old (I need R 4.4.0 or more).
I tried: module load GCC/12.3.0
module load GCCcore/12.3.0 module load UDUNITS/2.2.28 module load OpenMPI/4.1.5 module load GDAL/3.7.1 module load PROJ/9.2.0
module load R/4.3.2
and module load GCC/13.3.0 module load R/4.4.2
With R 4.4.2: terra package (required for biomod2) won’t install.
Does anyone know a recent R version and module setup on Baobab that allows installing terra package?
Thank you for your help 
Hi, I don’t use R. However, a generic method to install software not available on the cluster is by using apptainer (previously singularity). There are official Docker images for R available. A singularity image can be build from a docker image. If needed, the docker image can be customized using a Dockerfile.
In summary apptainer allows to create your own environment installing all the software you want. By default your home directory will be mounted.
Dear @Sandrine.Ranger
The problem is probably due to the fact that you are changing R versions: the packages you installed through R are in the Rpackages directory in your home directory.
If you install from R/4.3.2 and then switch to R/4.4.2, there will be conflicts and you will need to recompile the packages. The cleanest solution is to delete all the contents of Rpackages and start again.
I see that you have a lot of R packages in Rpackages: many of them are available in the R-bundle-CRAN/2024.11 module. I suggest you load it, as you will have much less to compile.
If compilation is too slow, I suggest you compile on a computing node.
