installing mixOmics in R within Anaconda

[Former user]

Hi, I've been trying to install mixOmics now for quite some time and seem to be getting nowhere. I have Mac OSX 10.13.1, the latest Anaconda version, Jupyter and R and RStudio versions. I've been following numerous guides online, recently following this guide: http://mixomics.org/wp-content/uploads/2016/09/Installation-guide-for-R-and-RStudio.pdf but still with no luck. My ultimate aim is to get mixOmics running within a Jupyter workflow with an R kernel running. I'm currently getting the following errors when trying to install from within RStudio. I've also tried installing the dependencies individually, with no luck. Thank you for any help you can provide.

Regards

Brett

install.packages("mixOmics") also installing the dependencies ‘igraph’, ‘rgl’

trying URL 'https://cran.rstudio.com/src/contrib/igraph_1.2.1.tar.gz' Content type 'application/x-gzip' length 2703426 bytes (2.6 MB) ================================================== downloaded 2.6 MB

trying URL 'https://cran.rstudio.com/src/contrib/rgl_0.99.16.tar.gz' Content type 'application/x-gzip' length 3058794 bytes (2.9 MB) ================================================== downloaded 2.9 MB

trying URL 'https://cran.rstudio.com/src/contrib/mixOmics_6.3.1.tar.gz' Content type 'application/x-gzip' length 3416045 bytes (3.3 MB) ================================================== downloaded 3.3 MB

• installing source package ‘igraph’ ... ** package ‘igraph’ successfully unpacked and MD5 sums checked checking for gcc... x86_64-apple-darwin13.4.0-clang checking whether the C compiler works... no configure: error: in /private/var/folders/0w/h96n74f572zd5kcl4c91xxbr3k14tm/T/RtmpPg45S0/R.INSTALLfb572b5c24c2/igraph': configure: error: C compiler cannot create executables Seeconfig.log' for more details ERROR: configuration failed for package ‘igraph’
• removing ‘/Users/bschapma/anaconda2/envs/bioconda/lib/R/library/igraph’ Warning in install.packages : installation of package ‘igraph’ had non-zero exit status
• installing source package ‘rgl’ ... ** package ‘rgl’ successfully unpacked and MD5 sums checked checking for gcc... x86_64-apple-darwin13.4.0-clang checking whether the C compiler works... no configure: error: in /private/var/folders/0w/h96n74f572zd5kcl4c91xxbr3k14tm/T/RtmpL6Mzh4/R.INSTALLfd27349cdad8/rgl': configure: error: C compiler cannot create executables Seeconfig.log' for more details ERROR: configuration failed for package ‘rgl’
• removing ‘/Users/bschapma/anaconda2/envs/bioconda/lib/R/library/rgl’ Warning in install.packages : installation of package ‘rgl’ had non-zero exit status ERROR: dependencies ‘igraph’, ‘rgl’ are not available for package ‘mixOmics’
• removing ‘/Users/bschapma/anaconda2/envs/bioconda/lib/R/library/mixOmics’ Warning in install.packages : installation of package ‘mixOmics’ had non-zero exit status

The downloaded source packages are in ‘/private/var/folders/0w/h96n74f572zd5kcl4c91xxbr3k14tm/T/RtmpTRv9vW/downloaded_packages’ Updating HTML index of packages in '.Library' Making 'packages.html' ... done

[Florian Rohart]

Hi Brett,

As you mentioned, the problem seems to be with the installation of igraph and rgl, not really with mixOmics. We have seen a lot of installation errors with those two packages, but not the one you encounter here, which seems to be a missing link between your C compiler and R. I have no real ideas about how to solve that problem though, do you have gcc installed with anaconda? maybe the command line tools as well? maybe https://stackoverflow.com/questions/10357804/configure-error-c-compiler-cannot-create-executables can help?

[Brett Chapman]

Hi Florian

I eventually got mixOmics working in Jupyter, but its not a very elegant work around. I had to install the package version of R from CRAN (https://cloud.r-project.org/) and then install rpy2 without dependencies and manually include all other dependencies except R, to avoid it reinstalling the Anaconda version of R. As long as the version of R I have in my environment is the system wide version and not the Anaconda version it seems to work ok.

We would like to be able to install mixOmics in a Jupyter environment and its native R without having to do this, as we would like to have a slick installation procedure for others to use when sharing our workflows. Is there any chance that mixOmics could be updated in the future to work with the Anaconda version of R? All of our other workflows are in Python, and running mixOmics through rpy2 would be quite powerful, using R and Python in parallel. Thanks.

Brett