[BioC] BioConductor Deployment at a large site, suggestions?

Brian K Smith brian at stat.ohio-state.edu
Wed Dec 1 16:48:21 CET 2004

>Your answer most likely would be solved by one of the sections in the FAQ,
>"Downloading All Packages From A Repository".

Unfortunately, I had looked at the FAQ, and it didn't help much,
probably due to my unfamiliarity with the package.  The list gets me
the directories for reposTools, most of which are permission denied when
trying to browse and d/l the .tar.gz directly.

It also doesn't tell me all the packages and dependencies needed for the
"default" install of affy, cdna, and exprs.  Thus every time I upgrade the
package it's far more work than any other R package I've ever encountered,
as I have to figure out what all I need to install in what order.  Thus,
between that and not wanting to hit bioconductor.org 100 times for 100
machines, I upgrade BioConductor rarely.

Perhaps if I was an R guru this would be clear, but I'm a sys admin so
I'm between the simple single user who just wants getBioC() and be done,
and someone who wants their own repository of all the possible packages.
I know just enough to install and basic test R and packages, though I
would love to learn more R, but due to workload, I just don't have time.

I would like an ordered list of packages and dependencies for the default
install into a typical R install.  Thus, I could then script this so
when I need to kickstart install a machine the entire process gets done
in 30 minutes and I have a machine that looked the same as it did before.

Though the smaller ordered R CMD INSTALL list would be ideal, I think
the reposTools and getBioC combined with an R BATCH may work.  I'll
look into it and probably contact you or Robert off list.

I can't imagine that I'm the only sys admin with this issue.  Perhaps
when done I can write a HOWTO and post it back to the list.


More information about the Bioconductor mailing list