M. K. wrote: > Dear Sir, > > Our unix cluster have R version 1.9.0. When I type > /source("http://www.bioconductor.org/getBioC.R")/ and /getBioC()/, I > get the following message: "You are currently running R version 1.9.0, > however R version 2.0.0 is required". Our system-add-min is not willing > to install the newest version (2.0.1) of the R, so is it possible to > install the bioconductor packages wihtout changing the R version ? > > For example, can I delete the version check in getBioC.R-code or > manually install the packages ? > Hi, the best place to ask these questions is the bioconductor mailing list. Do not simply elimate the version check. It is there for a reason and Bioconductor release 1.5 will not run on anything before R 2.0.0! You could try to revert back to release 1.4 of Bioconductor. I am not sure how to do this best, maybe by making some changes to the getbioC.R script, or by manually getting all required packages from http://www.bioconductor.org/repository/release1.4/package/html/index.h tml But the recommended solution is: Install your own R in your home directory! This is has many advantages, and it is not difficult. Simply get R-2.0.1.tar.gz, untar, run './configure' and 'make'. More detailed instruction are on the Biocondutor and R-project websites. For some options you may need to install additional libraries (e.g. on Suse "readline-devel" for nicer command-line editing, or libpng for PNG graphics, but these are second-order cosmetics.) If you want to seriously use R/Bioconductor, this is much better than relying on an unhelpful sysad. Also, in principle installing R and R packages as a user is more secure than as root! Although I have never heard of any cases, a malevolent package author could to bad things when you let the package installation run with superuser permissions... Best regards Wolfgang ------------------------------------- Wolfgang Huber European Bioinformatics Institute European Molecular Biology Laboratory Cambridge CB10 1SD England Phone: +44 1223 494642 Fax: +44 1223 494486 Http: www.ebi.ac.uk/huber

Wolfgang Huber <huber@ebi.ac.uk> wrote: > . . . > You could try to revert back to release 1.4 of Bioconductor. I am not > sure how to do this best, maybe by making some changes to the getbioC.R > script, or by manually getting all required packages from > http://www.bioconductor.org/repository/release1.4/package/html/index .html Having been down this road myself, I can confirm that the above will work, but with one major deficiency: I have been unable to get hold of most of the metadata packages which go with release 1.4. In the end, I gave up on this approach and have upgraded to R-2.0.1 & BioC-1.5, etc. > . . . > If you want to seriously use R/Bioconductor, this is much better than > relying on an unhelpful sysad. Also, in principle installing R and R > packages as a user is more secure than as root! Although I have never > heard of any cases, a malevolent package author could to bad things when > you let the package installation run with superuser permissions... Please pardon as I veer slightly farther off-topic here: Being a sysadmin myself (and not a BioC user), I'll mention that this security concern can easily be dealt with by a number of techniques. And I'll add that you might be better served if you advocated for your organization's getting more sysadmin's to take care of your research tools for you, instead of having each and every researcher go off and spend hours of their time learning the nuts and bolts of compilers and operating systems. Unless of course you enjoy doing that sort of thing and have someone willing to pay you to do it (:-). Regards, -- Marion Hakanson