[BioC] Need help on understanding this error message

Mon Apr 17 08:00:07 CEST 2006

The following is happening: you have read the CEL files into an  
AffyBatch object. The Cel files comes from the chip type "mu11ksuba".  
When you type
   dat
at the prompt, the AffyBatch object is printed to the screen. In  
order to print the object you need the CDF environment for the chip  
installed. BioC detects this and goes online to fetch the cdf  
environment (called mu11ksubacdf). It then tries to install the  
package but fails because the library is locked. This may be due to  
another process affecting the library at the same time, or perhaps  
you quit during an installation. The library you are trying to  
install into is the system wide library, which you may not have write  
access to.

Solution: tell your system admin to install the mu11ksubacdf package.

/Kasper

On Apr 14, 2006, at 11:19 AM, Fang Cheng wrote:

> Dear all,
>
> I was trying to load the .CEL files into R. I called affy() function.
> Did not get any problem in loading it to object "dat", but when I  
> tried
> to see what's in dat, I got the following error message, which I could
> not understand. Could someone give me some advices? I appreciate it in
> advance!
> ============================================================
>> dat
> trying URL
> 'http://bioconductor.org/packages/data/annotation/1.7/src/contrib/ 
> mu11ksubacdf_1.10.0.tar.gz'
> Content type 'application/x-gzip' length 621868 bytes
> opened URL
> ==================================================
> downloaded 607Kb
>
> mkdir: cannot create directory `/usr/local/lib/R/site-library/00LOCK':
> Permission denied
> ERROR: failed to lock directory '/usr/local/lib/R/site-library' for
> modifying
>
> The downloaded packages are in
>         /tmp/RtmpuEzo5X/downloaded_packages
> Error in getCdfInfo(object) : Could not obtain CDF environment,  
> problems
> encountered:
> Specified environment does not contain Mu11KsubA
> Library - package mu11ksubacdf not installed
> Data for package affy did not contain mu11ksubacdf
> Library - package mu11ksubacdf not installed
> In addition: Warning messages:
> 1: installation of package 'mu11ksubacdf' had non-zero exit status in:
> install.packages(cdfname, lib = lib, repos = Biobase:::biocReposList 
> (),
> 2: cannot create HTML package index in: tools:::unix.packages.html 
> (.Library)
> AffyBatch object
> size of arrays=534x534 features (22284 kb)
> cdf=Mu11KsubA (??? affyids)
> number of samples=10
> trying URL
> 'http://bioconductor.org/packages/data/annotation/1.7/src/contrib/ 
> mu11ksubacdf_1.10.0.tar.gz'
> Content type 'application/x-gzip' length 621868 bytes
> opened URL
> ==================================================
> downloaded 607Kb
>
> mkdir: cannot create directory `/usr/local/lib/R/site-library/00LOCK':
> Permission denied
> ERROR: failed to lock directory '/usr/local/lib/R/site-library' for
> modifying
>
> The downloaded packages are in
>         /tmp/RtmpuEzo5X/downloaded_packages
> Error in getCdfInfo(object) : Could not obtain CDF environment,  
> problems
> encountered:
> Specified environment does not contain Mu11KsubA
> Library - package mu11ksubacdf not installed
> Data for package affy did not contain mu11ksubacdf
> Library - package mu11ksubacdf not installed
> In addition: Warning messages:
> 1: missing cdf environment ! in: show(list())
> 2: installation of package 'mu11ksubacdf' had non-zero exit status in:
> install.packages(cdfname, lib = lib, repos = Biobase:::biocReposList 
> (),
> 3: cannot create HTML package index in: tools:::unix.packages.html 
> (.Library)
> ===============================================
>
> Thank you,
> Fang
>
>
> ----- Original Message -----
> From: Gordon Smyth <smyth at wehi.edu.au>
> Date: Wednesday, April 12, 2006 7:00 pm
> Subject: [BioC]  LIMMA: testing for batch effects
>
>>
>>> Date: Tue, 11 Apr 2006 15:08:55 +0000
>>> From: Adaikalavan Ramasamy <ramasamy at cancer.org.uk>
>>> Subject: [BioC] LIMMA: testing for batch effects
>>> To: BioConductor mailing list <bioconductor at stat.math.ethz.ch>
>>>
>>> Dear all,
>>>
>>> We have 63 arrays that are either diseased or normal. We wish to
>> select>genes adjusting for two covariates : gender (male/female) and
>>> experimental batch (one/two/three).
>>>
>>>> From biological knowledge, we expect the batch effect to be
>> significant>but we wish to quantify it numerically. Here is a way
>> that we tried and
>>> the problems we faced. We searched the archives without much  
>>> success.
>>>
>>>
>>>  library(limma)  # version 2.3.3
>>>  dd  <- model.matrix( ~ disease + gender + batch, cl )
>>>  head( dd )
>>>        (Intercept) diseasenormal gendermale batchtwo batchthree
>>>  2405            1             1          0        0          0
>>>  2408            1             1          0        0          0
>>>  2410            1             1          0        0          0
>>>  GER15           1             0          0        0          0
>>>  GER20           1             0          0        0          1
>>>  GER22           1             0          0        0          0
>>>
>>>  fit  <- lmFit(mds.rma, dd)
>>>
>>>
>>> 1) Is the following the correct way of setting up contrasts to
>> find the
>>> three pairwise comparison between batches ?
>>>
>>>  contr.m <- cbind( Batch2minus1=c(0,0,0,1,0),
>> Batch2minus1=c(0,0,0,0,1),>
>> Batch3minus2=c(0,0,0,-1,1) )
>>>
>>>
>>> 2) From reading the LIMMA user guide, we think decideTests()  
>>> could be
>>> potentially useful.
>>>
>>>  fit2 <- eBayes( contrasts.fit( fit, contr.m ) )
>>>  a <- decideTests( fit2, method="global" )
>>>  summary(a)
>>>     Batch2minus1 Batch2minus1 Batch3minus2
>>>  -1        15151        13838         4919
>>>   0        26574        30561        46703
>>>   1        12950        10276         3053
>>>
>>> Can we say that somewhere between 8000 - 27000 genes are affected
>> by at
>>> least one batch. Or is there a nicer/proper way of explaining this
>> to a
>>> biologist.
>>>
>>>
>>> 3) Ideally we would like to use method="nestedF" as suggested but
>> we get
>>> the following error message. Can anyone explain what might
>> possibly be
>>> going wrong.
>>>
>>>  b <- decideTests( fit2, method="nestedF" )
>>>  Error in if (crossprod(crossprod(Q, x)) > qF[i]) { :
>>>         missing value where TRUE/FALSE needed
>>
>> This was a bug that persisted for a short time after the built-in R
>> function p.adjust() changed from "fdr" to "BH" in R 2.2.0. The
>> change-log entry for limma 2.3.6 reads:
>>
>> 6 November 2005: limma 2.3.6
>>
>> - decideTests() failed with method="nestedF",adjust.method="BH",
>> now fixed
>>
>> Regards
>> Gordon
>>
>>
>>> Any hints will be much appreciated.
>>>
>>> Regards,
>>>
>>> --
>>> Adaikalavan Ramasamy                    ramasamy at cancer.org.uk
>>> Centre for Statistics in Medicine       http://www.csm-oxford.org/
>>> Wolfson College Annexe
>> http://www.stats.ox.ac.uk/~ramasamy/>Linton Road
>>         Tel : 01865 284 408
>>> Oxford OX2 6UD                          Fax : 01865 284 424
>>
>> _______________________________________________
>> Bioconductor mailing list
>> Bioconductor at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>> Search the archives:
>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/ 
> gmane.science.biology.informatics.conductor