[BioC] xps Error in exonLevel(exonlevel, chiptype) : invalid argument ‘exonlevel’

[cstrato]

Dear Dick

When you look at the Affymetrix libraryfile "MoGene-1_0-st-v1.r3.mps" 
used for ExpressionConsole you will see that it contains 35557 entries, 
while the Affymetrix annotation file 
"MoGene-1_0-st-v1.na25.mm9.transcript.csv" contains 35567 entries. 
The annotation file contains in column "category" following categories:
main, control->affx, control->bgp->antigenomic, normgene->exon, 
normgene->intron

Since "control->bgp->antigenomic, normgene->exon, normgene->intron" have 
no annotation, I do not use these probesets, which explains the 
difference to ExpressionConsole.

Even if I would use exactly the same probesets as the ExpressionConsole, 
the results would be different, since ExpressionConsole uses only 
"Sketch-Quantile" normalization to avoid memory  problems, and not the 
usual quantile normalization.

If you want to test if my implementation of the RMA algorithm is 
correct, please use e.g. HG-U133_Plus_2 data. You will see that you will 
get the same results as the "affy" package.

Best regards
Christian


Dick Beyer wrote:
> Hi Christian,
>
> Is there a choice of values to the agruments for rma that gets the 
> same results as Affy Expression Console?  I notice that from Affy 
> Expression Console I get an output with 35557 rows, but with either of 
> these choices:
>
> data.rma7 <- rma(data.laspada, 
> "laspadaCMRapRMA7",tmpdir="",background="antigenomic",normalize=TRUE, 
> verbose=TRUE,exonlevel="affx+core")
> data.rma8 <- rma(data.laspada, 
> "laspadaCMRapRMA8",tmpdir="",background="antigenomic",normalize=TRUE, 
> verbose=TRUE,exonlevel="all")
>
> then validData(data.rma7) or validData(data.rma8) gives me a matrix 
> with just 28837 rows.
>
> Thanks very much,
> Dick
> ******************************************************************************* 
>
> Richard P. Beyer, Ph.D.    University of Washington
> Tel.:(206) 616 7378    Env. & Occ. Health Sci. , Box 354695
> Fax: (206) 685 4696    4225 Roosevelt Way NE, # 100
>             Seattle, WA 98105-6099
> http://depts.washington.edu/ceeh/ServiceCores/FC5/FC5.html
> http://staff.washington.edu/~dbeyer
> ******************************************************************************* 
>
>
> On Fri, 9 May 2008, cstrato wrote:
>
>> Dear Dick
>>
>> I am glad to hear that you could now run rma.
>>
>> Please feel free to inform me about any potential problems, omissions 
>> in the help file, or potential improvements.
>>
>> Best regards
>> Christian
>>
>>
>> Dick Beyer wrote:
>>> Hi Christian,
>>>
>>> Thanks for your good help.  I'm a repeat offender when it comes to 
>>> not reading all the way through examples.  Thanks for your patience.
>>>
>>> Using background="antigenomic" made my rma call work just fine.
>>>
>>> Cheers,
>>> Dick
>>> ******************************************************************************* 
>>> Richard P. Beyer, Ph.D.    University of Washington
>>> Tel.:(206) 616 7378    Env. & Occ. Health Sci. , Box 354695
>>> Fax: (206) 685 4696    4225 Roosevelt Way NE, # 100
>>>             Seattle, WA 98105-6099
>>> http://depts.washington.edu/ceeh/ServiceCores/FC5/FC5.html
>>> http://staff.washington.edu/~dbeyer
>>> ******************************************************************************* 
>>> On Fri, 9 May 2008, cstrato wrote:
>>>
>>>> Dear Dick
>>>>
>>>> Thank you for your tests, which make it clear to me that I need to 
>>>> improve the help files.
>>>>
>>>> Since MoGene arrays as well as exon arrays have no MM, you need to 
>>>> set: background="antigenomic"
>>>>
>>>> Please have a look at directory "xps/examples" which contains files 
>>>> "script4xps.R" and "script4exon.R":
>>>> In these scripts I give numerous examples which you can copy/paste 
>>>> and adapt to your needs, especially I describe how to use the 
>>>> Affymetrix human tissue dataset with arrays HG-U133_Plus_2, 
>>>> HuGene-1_0-st-v1.r3 and HuEx-1_0-st-v2.r2.
>>>>
>>>> For HuGene the rma example is as follows:
>>>> data.rma <- 
>>>> rma(data.genome,"HuGeneMixRMAMetacore",filedir=datdir,tmpdir="",
>>>>                 
>>>> background="antigenomic",normalize=T,exonlevel="metacore+affx")
>>>>
>>>> Please let me know if these informations could help you solve your 
>>>> problem.
>>>>
>>>> Best regards
>>>> Christian
>>>>
>>>>
>>>> Dick Beyer wrote:
>>>>> Hi Christian,
>>>>>
>>>>> Thanks for pointing this out.  I tried several of the combinations 
>>>>> listed in the documentation for rma.  However, I now get a 
>>>>> different error message:
>>>>>
>>>>>> data.rma1 <- 
>>>>>> rma(data.laspada,"laspadaCMRapRMA1",tmpdir="",background="pmonly",normalize=TRUE, 
>>>>>> verbose=TRUE,exonlevel="core")
>>>>> Error in .local(object, ...) : error in function ‘PreprocessRMA’
>>>>>> data.rma2 <- 
>>>>>> rma(data.laspada,"laspadaCMRapRMA2",tmpdir="",background="pmonly",normalize=TRUE, 
>>>>>> verbose=TRUE,exonlevel="metacore")
>>>>> Error in .local(object, ...) : error in function ‘PreprocessRMA’
>>>>>> data.rma3 <- 
>>>>>> rma(data.laspada,"laspadaCMRapRMA3",tmpdir="",background="pmonly",normalize=TRUE, 
>>>>>> verbose=TRUE,exonlevel="affx")
>>>>> Error in .local(object, ...) : error in function ‘PreprocessRMA’
>>>>>> data.rma4 <- 
>>>>>> rma(data.laspada,"laspadaCMRapRMA4",tmpdir="",background="pmonly",normalize=TRUE, 
>>>>>> verbose=TRUE,exonlevel="affx+core")
>>>>> Error in .local(object, ...) : error in function ‘PreprocessRMA’
>>>>>> data.rma5 <- 
>>>>>> rma(data.laspada,"laspadaCMRapRMA5",tmpdir="",background="pmonly",normalize=TRUE, 
>>>>>> verbose=TRUE,exonlevel="all")
>>>>> Error in .local(object, ...) : error in function ‘PreprocessRMA’
>>>>>> data.rma6 <- 
>>>>>> rma(data.laspada,"laspadaCMRapRMA6",tmpdir="",background="pmonly",normalize=TRUE, 
>>>>>> verbose=TRUE,exonlevel="")
>>>>> Error in exonLevel(exonlevel, chiptype) : invalid argument 
>>>>> ‘exonlevel’
>>>>>
>>>>> Would you please suggest what I could do next?
>>>>> Thanks very much,
>>>>> Dick
>>>>> ******************************************************************************* 
>>>>> Richard P. Beyer, Ph.D.    University of Washington
>>>>> Tel.:(206) 616 7378    Env. & Occ. Health Sci. , Box 354695
>>>>> Fax: (206) 685 4696    4225 Roosevelt Way NE, # 100
>>>>>             Seattle, WA 98105-6099
>>>>> http://depts.washington.edu/ceeh/ServiceCores/FC5/FC5.html
>>>>> http://staff.washington.edu/~dbeyer
>>>>> ******************************************************************************* 
>>>>> On Fri, 9 May 2008, cstrato wrote: