[BioC] Administrative Rigths
Kasper Daniel Hansen
khansen at stat.Berkeley.EDU
Mon Apr 17 08:13:31 CEST 2006
On Apr 13, 2006, at 2:19 AM, marco_oliviero77 wrote:
> Dear BioC,
> I'm not a user of bioconductor, I am a system manager of a facility
> that will run R-Bioconductor for different users on a linux platform.
> They actually complain about the installation of metadata, as they
> dont have permissions to install them. Is there a way to give them
> partial permission to install these packages and further R packages
> without give them root controls?
Yes. There are many ways to do so. One way is to create a dedicated R
group (with write access). If you have a large set of inexperienced
users this may give problems as they may affect each other. Another
way is to make your users install their own libraries in $HOME.
Basically R has a list of libraries, exactly like the PATH env.
variable - this is called R_LIBS. The library list can also be set by
modifying Rprofile.site (this way you create a default for all users)
or $HOME/.Rprofile or $HOME/.Renviron (the path is set on a user-
level basis). as opposed to the PATH variable, the system wide
library is always in the library list (except if you really really
work hard to get rid of it :)
What I would do as a start is to identify the chips used by your
facility. My guess is that the list is rather limited. And then
install all the metadata associated with those chips. Most chips have
a chiptype, chiptypecdf, chiptypeprobe associated with them, eg.
You do this easiest by starting R as root (or however owns it) and do
R> biocLite(c("hgu95av", "hgu95av2cdf", "hgu95ac2probe"))
If your users have a specific need they can install special metadata
packages by setting their own R_LIBS. They should be educated on how
to do this anyway as you probably will see that you users constantly
wants top upgrade or try new things out.
Further info in the installation manual as well as the help pages for
Startup and R_LIBS (accessed on the web or by typing ?Startup or ?
R_LIBS at the R prompt) and the links in those pages.
So the possibilities are endless here.
Basically you need t
> thanks to all
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> Search the archives: http://news.gmane.org/
More information about the Bioconductor