[BioC] problem with loading DESeq after upgrading from R 2.14 to R 2.15

Mon Apr 23 23:23:20 CEST 2012

On 4/23/2012 5:17 PM, Dan Tenenbaum wrote:
> On Mon, Apr 23, 2012 at 2:11 PM, Yue Li<gorillayue at gmail.com>  wrote:
>> Any helps on not having to manually re-install the packages for each R
>> upgrade would be greatly appreciated! I followed this post but apparently
>> not be able to make the upgrade seamless ...
>>
> I echo Steve in saying this is not really recommended.
> For a given version of R, you should be using the appropriate version
> of Bioconductor.
>
> If it makes it easier, you can go into R 2.14 and get a list of the
> packages you have installed with
> pkgs<- installed.packages()[,"Package"]
>
> You can then take this list and use it as an argument to biocLite() in R 2.15.
> However, the list will include packages that are part of R so these
> will not install properly.

Or alternatively,

x <- installed.packages()
y <- row.names(x)[-grep("Part", x[,11])]

Then you can write that out and scan() it back into your new R version.

Best,

Jim

>
> Dan
>
>
>> https://stat.ethz.ch/pipermail/r-sig-mac/2008-August/005276.html
>>
>>
>> On 2012-04-23, at 4:54 PM, Dan Tenenbaum wrote:
>>
>> On Mon, Apr 23, 2012 at 1:47 PM, Yue Li<gorillayue at gmail.com>  wrote:
>>
>> Hi Dan,
>>
>>
>> By "all of the packages", I meant the packages installed for R 2.14 not all
>> of the packages from CRAN or Bioconductors.
>>
>>
>> They work fine under terminal as shown below and most work fine in R GUI
>> except for DESeq and DEXSeq ...
>>
>>
>> OK, thanks for the clarification.
>>
>> Dan
>>
>>
>>
>> Yue
>>
>>
>> On 2012-04-23, at 4:39 PM, Dan Tenenbaum wrote:
>>
>>
>> Hi Yue,
>>
>>
>> On Mon, Apr 23, 2012 at 1:10 PM, Yue Li<gorillayue at gmail.com>  wrote:
>>
>> Hi Steve,
>>
>>
>> Thanks for prompt reply! It's just that having to re-install all of the
>> packages for every upgrade of R is also a hassle, don't u think? But you're
>> right that installing it improperly is even disastrous like what I am
>> experiencing now.
>>
>>
>> When you say "all of the packages," what do you mean? You only need to
>>
>> install the packages that you work with. There is no requirement to
>>
>> install all of them.
>>
>>
>> Dan
>>
>>
>>
>>
>>
>> I wish in the future, someone could just make the upgrade of R as easy as
>> upgrading iTunes or something like that.
>>
>>
>> Yue
>>
>>
>>
>> On 2012-04-23, at 3:53 PM, Steve Lianoglou wrote:
>>
>>
>> Hi,
>>
>>
>> The way you are trying to maintain your libraries is generally frowned
>>
>> upon. You shouldn't be trying to use R-2.14 packages with your R-2.14
>>
>> install. You are exposing yourself to weird and hard to explain
>>
>> problems.
>>
>>
>> Look at the error message you are getting:
>>
>>
>> On Mon, Apr 23, 2012 at 3:36 PM, Yue Li<yueli at cs.toronto.edu>  wrote:
>>
>> [snip]
>>
>> Loading required package: locfit
>>
>> locfit 1.5-7     2012-03-22
>>
>> Error in dyn.load(file, DLLpath = DLLpath, ...) :
>>
>> unable to load shared object '/Library/RLib/RSQLite/libs/x86_64/RSQLite.so':
>>
>> dlopen(/Library/RLib/RSQLite/libs/x86_64/RSQLite.so, 6): Library not loaded:
>> /Library/Frameworks/R.framework/Versions/2.14/Resources/lib/libR.dylib
>>
>> Referenced from: /Library/RLib/RSQLite/libs/x86_64/RSQLite.so
>>
>> Reason: image not found
>>
>> Error: package/namespace load failed for ‘DESeq’
>>
>>
>> Your trying to load a package from your 2.14 repo ... I'm not sure why
>>
>> this works from the terminal and not the GUI (are you sure you are
>>
>> firing up 2.15 from the terminal?), but I'm also not too inclined to
>>
>> smoke this out, because ... well, you shouldn't be doing this ;-) and
>>
>> also it will likely take you less time to just install things "the
>>
>> normal way".
>>
>>
>> If you really have some compelling reason to have a custom library
>>
>> structure, I like the way the bioc folks suggest developers to keep
>>
>> bioc-devel package separate from bioc-release (now that R is on an
>>
>> annual release cycle), eg:
>>
>>
>> http://www.bioconductor.org/developers/useDevel/
>>
>>
>> You could imagine having R-2.14 and R-2.15 in separate dirs instead of
>>
>> 2.15-release and 2.15-devel if working w/ devel libraries isn't your
>>
>> cup of tea -- but note they still keep different versions different.
>>
>>
>> Anyway, the "normal" R install does this for you anyway on OS X, so
>>
>> dollars-to-donuts you'll have less hassles just going "the normal"
>>
>> route, and installing the package you need via biocLite again in your
>>
>> "clean" 2.15 library path will likely cost less time than trying to
>>
>> debug your custom library path woes.
>>
>>
>> HTH,
>>
>> -steve
>>
>>
>>
>> --
>>
>> Steve Lianoglou
>>
>> Graduate Student: Computational Systems Biology
>>
>> | Memorial Sloan-Kettering Cancer Center
>>
>> | Weill Medical College of Cornell University
>>
>> Contact Info: http://cbio.mskcc.org/~lianos/contact
>>
>>
>> _______________________________________________
>>
>> Bioconductor mailing list
>>
>> Bioconductor at r-project.org
>>
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>
>> Search the archives:
>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>
>>
>>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, M.S.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099