[BioC] Follow-up: recipe for installing and using Bioconductor with Mac OS X 10.3 and R GUI 1.9

Marc Salit salit at nist.gov
Thu Apr 22 22:08:41 CEST 2004


Patient Colleagues --

Further follow-up, thanks mostly to Robert Gentleman and Stefano  
Iacus...

It was suggested that I not live with the permissions open in the R  
framework (because of the security holes that leaves), and that I take  
another shot at installing the latest version of R GUI for the Mac and  
the packages as pre-compiled binaries. I dragged the permission-changed  
R.framework directory tree to the trash, emptied the trash, and started  
over with the fresh package.

Slightly longer story made short: this all worked. I installed the  
package from the CRAN website, and then used the menu option  
Packages|Bioconductor|Individual Packages|Ready-to-use Binary to  
install biobase, affy, and the widget tools. Then I ran the X windows  
server in Mac OS X (to accommodate the Tcl/tk stuff) before calling  
expresso with widget=TRUE. Viola.

Thanks for the help. Now to converge on some consistent, credible,  
coherent expression measures...

Best regards,
Marc Salit

--  
Dr. Marc Salit
Research Chemist
NIST
Chemical Science and Technology Laboratory
100 Bureau Drive, Stop 8391
Gaithersburg, MD 20899-8391
email: Salit domain NIST.gov
voice: +1 301 975-3646
fax:   +1 301 869-0413

On Apr 21, 2004, at 6:43 PM, Marc Salit wrote:

> Dear Colleagues --
>
> A little follow-up is in order -- I've got things going, more or less.  
> Mostly more. Thanks for the suggestions, which mostly got me on track.  
>  Bioconductor things seem to work in general, it was the affy package  
> I was having trouble with. I am still having problems with the  
> 'widget' tools, but have received some leads on that (problems with  
> tcltk).
>
> Ok, so I'm running R release version Version 1.9.0  (2004-04-12) , on  
> Mac OS X 10.3.3 on a PowerMac G5 2.0 GHz dual processor.
>
> Some things that I did in order to get going:
> 	1) I installed the Apple Mac OSX developer's tools, which are needed  
> to compile source.
>
> 	2) I changed owner and permissions on the R.Framework directory, and  
> everything underneath it to permit r/w access to all users. (From some  
> of the error messages I was receiving, it seemed that this was  
> necessary, though your mileage may vary...)
>
> 	3) I downloaded the development version of the affy package  
> (affy_1.4.23.tar.gz) and installed from the command line. The  
> 'released' version (affy_1.3.28.tar.gz) would not install this way, it  
> died while building help for normalize.invariantset (see below).
>
>   normalize.contrast                text    html    latex   example
>   normalize.invariantset            text    html    latex
>
> ERROR: command (e.g. \url) inside \link
> ERROR: building help failed for package 'affy'
> ** Removing  
> '/Library/Frameworks/R.framework/Versions/1.9.0/Resources/library/ 
> affy'
>
> And, my problems with tcltk:
>
> > package.manager()
> loading package: widgetTools
> Loading required package: tcltk
> Error in fun(...) : this isn't a Tk applicationcouldn't connect to  
> display ":0"
> Error: .onLoad failed in loadNamespace
> Error in f(libname, pkgname) : Package tcltk not able to be loaded!
> Error in library(pkgs[i], character.only = TRUE) :
> 	.First.lib failed
> > dchip <- expresso(widget=TRUE)
> Loading required package: tkWidgets
> Error in expresso(widget = TRUE) : library tkWidgets could not be  
> found !
> In addition: Warning message:
> There is no package called 'tkWidgets' in: library(package,  
> character.only = TRUE, logical = TRUE, warn.conflicts =  
> warn.conflicts,
> > package.manager()
> loading package: tcltk
> Error in fun(...) : this isn't a Tk applicationcouldn't connect to  
> display ":0"
> Error: .onLoad failed in loadNamespace
> Error in library(pkgs[i], character.only = TRUE) :
> 	package/namespace load failed
>
> Best regards,
> Marc Salit
>
> On Apr 13, 2004, at 5:15 PM, Marc Salit wrote:
>
>> Dear Colleagues --
>>
>> I've been trying unsuccessfully to use Bioconductor (mostly the affy  
>> package) on a Mac with OS X 10.3.3, running R with a GUI, either  
>> 1.8.1 or 1.9.0 beta. I've tried to install Bioconductor with  
>> getBioC.R, which doesn't work all the way, and also tried to use a  
>> shell command to install downloaded packages, a la R CMD INSTALL  
>> /path/to/pkg_version.tar.gz , which also doesn't work all the way  
>> (stops when it can't find a 'make' command in SHLIB.
>>
>> Suggestions? Caveats? Will this work? Shall I give up and 'switch'  
>> back to a PC???
>>
>> Cheers & TIA,
>> Marc Salit
>>
>> -- 
>> Dr. Marc Salit
>> Research Chemist
>> NIST
>> Chemical Science and Technology Laboratory
>> 100 Bureau Drive, Stop 8391
>> Gaithersburg, MD 20899-8391
>> email: Salit at NIST.gov
>> voice: +1 301 975-3646
>> fax:   +1 301 869-0413_______________________________________________
>> Bioconductor mailing list
>> Bioconductor at stat.math.ethz.ch
>> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 5364 bytes
Desc: not available
Url : https://www.stat.math.ethz.ch/pipermail/bioconductor/attachments/20040422/ea79ed40/attachment.bin


More information about the Bioconductor mailing list