edgeR installation error inside notebooks

[adeslatt]

This could be a bioconductor problem - Initially created an environment - sbas2 inside the terminal window -- loaded all required packages. When starting the differentialSplicingJunctionAnalysis.ipynb notebook, received an error when instantiating the libraries. non-zero status when loading edgeR.

Went to the command line. Ran R from the command line.

install.packages("BiocManager")
BiocManager::install("edgeR")
library("edgeR")

this now worked. Went back to the notebook and ran

library("edgeR")

and this worked.
@cgpu what I don't know or understand how the conda created environment is inherited within a notebook -- the environment proliferates under a different name -- I called my conda created environment sbas2 but inside the notebook it shows as sbas. I will try again with a non-numeric name.

I am suspicious that the error we saw with the conda install of libraries is real -- because the installation of both edgeR and of multtest when using BiocManager performs a recompile that appears to be mandated. When this is done within the sbas2 environment and then I load the library within the notebook -- there is not an error.