[BioC] affylmGUI segfault on "pd.ragene.1.0.st.v1"

James W. MacDonald jmacdon at med.umich.edu
Tue Jan 6 22:46:57 CET 2009


Hi Matt,

The problem as I see it is that affylmGUI is specifically designed to 
work with either the affy, gcrma or affyPLM packages and the S4 
structures produced by those packages. So while you could try to make it 
so an end user could load oligo after e.g., the affy package and thereby 
mask all of the exposed functions from affy so they could then use 
affylmGUI with oligo and fool it into using oligo's functions, I just 
don't see the point.

For instance, one could theoretically load affylmGUI, then oligo, and 
the rma() function from oligo would mask that from affy. Then when you 
try to normalize the data and affylmGUI calls rma() you would get the 
oligo version, which presumably would then work correctly with the data 
at hand.

However, I think it is preferable that the maintainer of affylmGUI would 
decide of his own volition to support the newer affy chips and directly 
incorporate the functionality rather than modifying oligo so an end user 
could theoretically trick the package into using oligo.

Best,

Jim

Matt Settles wrote:
> James,
> 
> Wouldn't that depend on how the information is accessed? If affylmgui 
> uses functions for ALL its data access from the cdf env, then I could 
> create functions in oligo that return the requested information in the 
> expected form regardless of the underlying data structure. If affylmgui 
> accesses the data directly with assumed data types then your absolutely 
> correct and any changes would need to be made by that group. I don't 
> know which is the case, but I would be very interested to know.
> 
> One of my goals with these chip types in oligo is to be able to provide 
> interface functions similar to what people are already used to with 
> affy,, when appropriate, and hence better interoperability with existing 
> packages.
> 
> James W. MacDonald wrote:
>> Hi Matt,
>>
>> Matt Settles wrote:
>>> Harry, Wolfgang and James,
>>>
>>> The exon and gene ST arrays do not use, or even have, a CDF file for 
>>> the array, I think a requirement for makecdfenv. These arrays use clf 
>>> and pgf files, hence the need for the oligo set of packages.
>>>
>>> You can try changing the name of pd.ragene.1.0.st.v to 
>>> ragene10stv1cdf then reinstalling. This should at least enable you to 
>>> load the package within affylmgui, there may be errors when affylmgui 
>>> tries to access the information in the package and I would be 
>>> interested in what they are. I can then work on updating oligo to add 
>>> an interface similar to what the affylmgui is expecting.
>>
>> I don't think that will be helpful. The pdInfo packages are based on 
>> an underlying SQLite database whereas the cdfenvs are environment 
>> based. Simply renaming something to look compatible and then trying to 
>> use it will not work.
>>
>> And affylmGUI is expecting to use the affy package (hence the name), 
>> so you won't be able to add anything to oligo that will help. Any 
>> changes will have to occur in affylmGUI (or oligolmGUI if Keith 
>> Satterley wants a new project ;-D).
>>
>> Best,
>>
>> Jim
>>
>>>
>>> Matt
>>>
>>>
>>> Wolfgang Huber wrote:
>>>>
>>>> Hi Harry & Matt
>>>>
>>>> re your second question, about the CDF package: the way the affy 
>>>> package works, it expects a package with name "ragene10stv1cdf", 
>>>> containing an R-environment of that same name, with the probeset 
>>>> definitions for that chip.
>>>>
>>>> Note that the CDF packages that the "affy" package uses are *not* 
>>>> the same as the platform design packages that the "oligo" package uses.
>>>> Essentially: affy = old, tried and tested, but only works for 
>>>> convential Affymetrix genechips for mRNA analysis with genechip 
>>>> using 3' probe sets.
>>>> oligo = newer, more general, also suitable for more recent chip 
>>>> types inclduding for genotyping..
>>>>
>>>> It seems that you want affy and CDF packages. The make.cdf.package 
>>>> function from the makecdfenv package is what you or Matt would want 
>>>> to use to build these.
>>>>
>>>> Jim McDonald (cc-ed) has written an overview over the software tools 
>>>> for analysing Affymetrix arrays, unfortunately I could not figure 
>>>> out how to find it from the Bioconductor webpage, so maybe he can 
>>>> help with that.
>>>>
>>>>
>>>> Best wishes
>>>>  Wolfgang
>>>>
>>>> Harry Mangalam ha scritto:
>>>>> Hi All,
>>>>>
>>>>> I've been trying to get affylmGUI to run on my laptop (3GB, Ubuntu 
>>>>> 8.04, R as below:
>>>>> =====
>>>>>> sessionInfo()
>>>>> R version 2.8.0 (2008-10-20)
>>>>> i686-pc-linux-gnu
>>>>>
>>>>> locale:
>>>>> LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=C;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C 
>>>>>
>>>>>
>>>>> attached base packages:
>>>>> [1] tools     tcltk     stats     graphics  grDevices utils     
>>>>> datasets
>>>>> [8] methods   base
>>>>>
>>>>> other attached packages:
>>>>> [1] affylmGUI_1.16.0     affy_1.16.0          preprocessCore_1.0.0
>>>>> [4] affyio_1.6.1         Biobase_2.2.1        limma_2.16.3
>>>>> =====
>>>>> The commands to start affylmGUI are:
>>>>> library("pd.ragene.1.0.st.v1") # the array set, from Matt Settles
>>>>> library("oligo")
>>>>> library("affylmGUI")
>>>>> affylmGUI()
>>>>>
>>>>> when I try to load data, whether my own (or the estrogen data set), 
>>>>> it segfaults at the same place:
>>>>> After I select a working directory, and select the targets file 
>>>>> (included below), [R churns for a few seconds], and give the new 
>>>>> data set a name, I can see that the data is 'available' via the 
>>>>> data tree in the left panel.
>>>>>
>>>>> At the point where I try to view the RNA targets via Menu -> RNA 
>>>>> Targets -> RNA Targets, and release the mouse button, it segfaults, 
>>>>> leaving the traceback:
>>>>>
>>>>> http://moo.nac.uci.edu/~hjm/affylmGUI.traceback
>>>>>
>>>>> As I said above, this is the case if it's my data or the estrogen 
>>>>> data set.
>>>>>
>>>>> If I try to bypass that step and go direct to normalization (with 
>>>>> my data), it will eventually emit the error that I don't have the 
>>>>> correct CDF environment:
>>>>>
>>>>> ==== Error message ====
>>>>> Error in getCdfInfo(object): Could not obtain CDF environment, 
>>>>> problems encountered: Specified environment does not contain 
>>>>> RaGene-1_0-st-v1 Library - package ragene10stv1cdf not installed. 
>>>>> Data for package affy did not contain ragene10stv1cdf Bioconductor 
>>>>> - ragene10stv1cdf not available
>>>>> =======================
>>>>>
>>>>> This despite loading without error the correct (from Matt Settles) 
>>>>> library beforehand. > library("pd.ragene.1.0.st.v1")
>>>>> Does the spelling variant muck it up? Best wishes,
>>>>> harry
>>>>>
>>>>>
>>>>> ==== targets.txt (whitespaces are TABS) ====
>>>>>
>>>>> Name    FileName        Target
>>>>> AB.1    0908F-02_AB_1.CEL       AB_Only
>>>>> AB.2    0908F-02_AB_2.CEL       AB_Only
>>>>> AB.3    0908F-02_AB_3.CEL       AB_Only
>>>>> CIQ.1   0908F-02_CIQ_1.CEL      CIQ_Only
>>>>> CIQ.2   0908F-02_CIQ_2.CEL      CIQ_Only
>>>>> CIQ.3   0908F-02_CIQ_3.CEL      CIQ_Only
>>>>> CIQ+AB.1        0908F-02_CIQ_AB_1.CEL   CIQ_AB_CoTreatment
>>>>> CIQ+AB.2        0908F-02_CIQ_AB_2.CEL   CIQ_AB_CoTreatment
>>>>> CIQ+AB.3        0908F-02_CIQ_AB_3.CEL   CIQ_AB_CoTreatment
>>>>> Untreated.1     0908F-02_Untreated_1.CEL        CONTROL
>>>>> Untreated.2     0908F-02_Untreated_2.CEL        CONTROL
>>>>> Untreated.3     0908F-02_Untreated_3.CEL        CONTROL
>>>>>
>>>>
>>

-- 
James W. MacDonald, M.S.
Biostatistician
Hildebrandt Lab
8220D MSRB III
1150 W. Medical Center Drive
Ann Arbor MI 48109-5646
734-936-8662



More information about the Bioconductor mailing list