[BioC] snapCGH fails in runBioHMM

[J-C. Marioni]

Dear Ramon,

Sorry about the delay in responding to your messages - I was away last 
week. I suspect that some of the problems are due to the Namespace - I'll 
work on this today and upload a new version of the package later.

Apologies for any inconvenince.

John

On Tue, 12 May 2009, Ramon Diaz-Uriarte wrote:

> Dear Wolfgang,
>
> On Tuesday 12 May 2009 16:52:30 Wolfgang Raffelsberger wrote:
>> Hi Ramon,
>>
>> just to say that I posted another message concerning an error with
>> snapCGH package on this list on may 4th, but got no reply.
>
> Yes, I am expecting something similar. For my own particualr problem (with
> package ADaCGH) I've just surrounded all calls to snapCGH functions with
> a "try", so the help examples at least can run, and I return to the user the
> error message that results from snapCGH.
>
>
>> I my case it was one particular array that for some reason caused
>> runBioHMM() to stall.  I found a workaround, by just inverting R & G
>> channel for the array causing the trouble and runBioHMM() ran fine (of
>> course after segmentation I had to re-flip R&G for the array in question).
>> Besides, I had and "fixed" another problem with runGLAD() that wasn't
>> expecting snapCGH objects ...
>>
>> But I haven't tried the snapCGH v.1.12  yet ...
>
> As far as I can tell, there are no differences whatsoever in the code itself
> between versions 1.11 and 1.12 (or between 1.12 and 1.13). 1.12 and 1.13 seem
> identical except for their version number.  1.12 and 1.11 differ in a few
> changes in the documentation (formatting changes and fixes of types) and in
> the addition of the (buggy) NAMESPACE in 1.12.
>
> Best,
>
>
> R.
>
>
>
>>
>> cheers,
>> Wolfgang
>>
>> Ramon Diaz-Uriarte a écrit :
>>> Dear All,
>>>
>>> snapCGH (v. 1.12 in BioC 2.4 for sure; most likely v. 1.13 in BioC 2.5
>>> too) fails with the error
>>>
>>> C symbol name "runNelderMead" not in DLL for package "snapCGH"
>>>
>>> The problem is that the NAMESPACE file is not loading the required code.
>>> Sure enough, if one loads the .so file manually (e.g.,
>>> via "dyn.load('path-to-library/snapCGH.so')") things work.
>>>
>>>
>>> So, while the package gets fixed, what would be the orthodox way to get
>>> packages that depend on snapCGH to work?
>>>
>>>
>>> Best,
>>>
>>> R
>>>
>>> ************************
>>>
>>> Two reproducible examples:
>>>
>>> a) Using function fit.model (which is docummented but not exported in the
>>> NAMESPACE, which is available since v. 1.12):
>>>
>>> ydat <- matrix(rnorm(100), ncol = 1)
>>> datainfo <- data.frame(Name = 1:100, Chrom = rep(1, 100), Position =
>>> 1:100)
>>>
>>> snapCGH:::fit.model(sample = 1, chrom = 1, dat = ydat,
>>>                     datainfo = datainfo)
>>>
>>>
>>>
>>> b) Using the code in the vignette (the vignette does not fail, as the
>>> crucial chunck has "eval=FALSE", and the help for runBioHMM has no
>>> runnable examples):
>>>
>>> datadir <- system.file("testdata", package="snapCGH")
>>> targets <- readTargets("targets.txt", path=datadir)
>>> RG1 <- read.maimages(targets$FileName, path=datadir, source = "genepix")
>>> RG1 <- read.clonesinfo("cloneinfo.txt", RG1, path=datadir)
>>> RG1$printer <- getLayout(RG1$genes)
>>> types <- readSpotTypes("SpotTypes.txt", path=datadir)
>>> RG1$genes$Status <- controlStatus(types, RG1)
>>> RG1$design <- c(-1,-1)
>>> RG2 <- backgroundCorrect(RG1, method="minimum")
>>> MA <- normalizeWithinArrays(RG2, method="median")
>>> MA2 <- processCGH(MA,method.of.averaging=mean, ID = "ID")
>>> SegInfo.Bio <- runBioHMM(MA2)
>>>
>>>> version
>>>
>>>                _
>>> platform       x86_64-unknown-linux-gnu
>>> arch           x86_64
>>> os             linux-gnu
>>> system         x86_64, linux-gnu
>>> status         Patched
>>> major          2
>>> minor          9.0
>>> year           2009
>>> month          05
>>> day            11
>>> svn rev        48528
>>> language       R
>>> version.string R version 2.9.0 Patched (2009-05-11 r48528)
>>>
>>>> packageDescription("snapCGH")
>>>
>>> Package: snapCGH
>>> Title: Segmentation, normalisation and processing of aCGH data.
>>> Version: 1.12.0
>>> Date: 2000-04-07
>>> Author: Mike L. Smith, John C. Marioni, Steven McKinney, Thomas
>>>         Hardcastle, Natalie P. Thorne
>>> Description: Methods for segmenting, normalising and processing aCGH
>>>         data; including plotting functions for visualising raw and
>>>         segmented data for individual and multiple arrays.
>>> biocViews: Microarray, DNACopyNumber, TwoChannel, Preprocessing
>>> Maintainer: John Marioni <marioni at uchicago.edu>
>>> Depends: limma,tilingArray,DNAcopy,GLAD,cluster,methods,aCGH
>>> License: GPL
>>> Packaged: 2009-04-21 09:40:04 UTC; biocbuild
>>> Built: R 2.9.0; x86_64-unknown-linux-gnu; 2009-05-12 10:31:17 UTC; unix
>>
>> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
>> Wolfgang Raffelsberger, PhD
>> Laboratoire de BioInformatique et Génomique Intégratives
>> CNRS UMR7104, IGBMC,
>> 1 rue Laurent Fries,  67404 Illkirch  Strasbourg,  France
>> Tel (+33) 388 65 3300         Fax (+33) 388 65 3276
>> wolfgang.raffelsberger (at) igbmc.fr
>
>
>
> -- 
> Ramón Díaz-Uriarte
> Biocomputing Programm
> Centro Nacional de Investigaciones Oncológicas (CNIO)
> (Spanish National Cancer Center)
> Melchor Fernández Almagro, 3
> 28029 Madrid (Spain)
> Fax: +-34-91-224-6972
> Phone: +-34-91-224-6900
>
> http://ligarto.org/rdiaz
>
>
>
>
> **NOTA DE CONFIDENCIALIDAD** Este correo electrónico, y ...{{dropped:3}}
>
> _______________________________________________
> 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
>