Question: Does anyone manage to install Monocle3 on Rstudio 3.6.1 (on Debian 9)?
0
gravatar for castravete2712
13 days ago by
castravete27120 wrote:

Hello,

The installation of Monocle3 on Rstudio is driving me nuts and I'm trying to install it for a couple of days, by different methods ... nothing works. Monocle, the previous release of Monocle3 get to be installed but has some troubles running and I would prefer to work with the latest version and fonctionalities.

I'm running on Debian9, on Rstudio. For now, the latest R version is installed and it's R 3.6.1. I can't manage to install the 3.6.2 but normally, Monocle3 should work on 3.6.1 too.

Everything is working, Bioconductor, the packages that need to be installed before (like SingleCellExperiment), leidenbase through devtools....

So I had to install manually a couple of packages, like speedglm, furr ...

rstudio monocle3 R • 129 views
ADD COMMENTlink modified 13 days ago • written 13 days ago by castravete27120

Install speedglm manually and then post the error message. Doesn't work is difficult to debug without further details.

ADD REPLYlink written 13 days ago by ATpoint29k

Well, manually, I manage to install speedglm.

I encounter now a new problem:

configure: error: in `/tmp/RtmpDZcah8/R.INSTALL6eed630b38f3/units':
configure: error: 
--------------------------------------------------------------------------------
  Configuration failed because libudunits2.so was not found. Try installing:
    * deb: libudunits2-dev (Debian, Ubuntu, ...)
    * rpm: udunits2-devel (Fedora, EPEL, ...)
    * brew: udunits (OSX)
  If udunits2 is already installed in a non-standard location, use:
    --configure-args='--with-udunits2-lib=/usr/local/lib'
  if the library was not found, and/or:
    --configure-args='--with-udunits2-include=/usr/include/udunits2'
  if the header was not found, replacing paths with appropriate values.
  You can alternatively set UDUNITS2_INCLUDE and UDUNITS2_LIBS manually.
--------------------------------------------------------------------------------

See `config.log' for more details
ERROR: configuration failed for package ‘units’
* removing ‘/home/INT/draia-nicolau.o/miniconda/lib/R/library/units’
Error: Failed to install 'monocle3' from GitHub:
  (converted from warning) installation of package ‘units’ had non-zero exit status

I installed libudunits2-dev but even if a restart R after that, but the error continues to appear.

ADD REPLYlink modified 13 days ago • written 13 days ago by castravete27120
0
gravatar for castravete2712
13 days ago by
castravete27120 wrote:

Found a solution finally:

Install it through conda, it works perfectly

ADD COMMENTlink written 13 days ago by castravete27120
2

While I am a big fan of conda for basically everything except R, you might run into trouble when mixing CRAN/BioC and conda-installed packages and dependencies. Consider trying to solve the issues you have and compile from source to avoid that. https://cole-trapnell-lab.github.io/monocle3/docs/installation/

ADD REPLYlink modified 13 days ago • written 13 days ago by ATpoint29k

Ah indeed, that could occur. I'll try again correcting the issues I have with the installation from source the. Thank you!

ADD REPLYlink modified 13 days ago • written 13 days ago by castravete27120

I already encounter some problems with the conda version of the installation, it charges the monocle3 library but don't recognize functions from it, like newCellDataSet.

ADD REPLYlink written 13 days ago by castravete27120
3

If using conda, you should use it for everything, especially where R is concerned. This means installing everything R-related with conda. In general, mixing conda and system-based installation is a recipe for problems.

ADD REPLYlink written 13 days ago by Jean-Karim Heriche21k

Well, I installed R through conda. So you say that I have to install all packages for R through conda, the Rstudio (that I also installed with conda) package installation is not recommended even if R is installed through conda?

ADD REPLYlink modified 12 days ago • written 12 days ago by castravete27120
1

Yes. If you install R with conda, also install the packages with conda, don't do it from R.

EDIT: Also pay attention to which channel(s) you're using. There may be different versions of packages in different channels that may not be compatible. I'd suggest to stick to the following channels: defaults, conda-forge and bioconda.

ADD REPLYlink modified 12 days ago • written 12 days ago by Jean-Karim Heriche21k

Okay, I didn't realize how much things can get tangled up. Thank you very much for the different advices, I'll be more careful

ADD REPLYlink written 12 days ago by castravete27120
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1457 users visited in the last hour