[BioC] localization of mm values in affybatch exprs matrix

[Kasper Daniel Hansen]

So I was a bit quick. It seems from Karin's post that she already has  
a CDF env.

Jim's statement that the CEL file is ordered PM then MM is probably  
right for most chips, but in general you can only be sure that the PM  
and the MM are spatially close. In general you should use the CDF  
information to link the pm/mm/(x,y) position together and you cannot  
a priori know what coordinate corresponds to in terms of pm/mm/probeset.

Kasper

On Jan 26, 2007, at 2:56 PM, Kasper Daniel Hansen wrote:

>
> On Jan 25, 2007, at 10:10 AM, James W. MacDonald wrote:
>
>> Hi Karin,
>>
>> Karin Lagesen wrote:
>>> I have a custom affy chip that I read into R using ReadAffy():
>>>
>>>> newdata = ReadAffy()
>>>> newdata
>>>
>>> AffyBatch object
>>> size of arrays=754x754 features (17777 kb)
>>> cdf=E_colia530222N (11378 affyids)
>>> number of samples=4
>>> number of genes=11378
>>> annotation=ecolia530222n
>>>
>>>
>>> I now want to look at different values in this object.
>>>
>>> For instance, some pm values:
>>>
>>>
>>>> pm(newdata)[1:5,]
>>>
>>>      chip1.CEL chip2.CEL chip3.CEL chip4.CEL
>>> [1,]    1855.0    2180.8    1444.0  2932.0
>>> [2,]    2812.0    3451.0    2276.5  3406.0
>>> [3,]    4162.3    4301.0    2996.0  5088.0
>>> [4,]    1608.5    1758.0    1123.0  1987.0
>>> [5,]    2290.0    3189.0    2474.5  2838.3
>>>
>>>
>>> I now also look at the values in the affybatch exprs matrix:
>>>
>>>
>>>> newdata at exprs[1:5,]
>>>
>>>      chip1.CEL chip2.CEL chip3.CEL chip4.CEL
>>> [1,]     942.0     776.0       281    1475
>>> [2,]   24422.0   26071.0      8914   21826
>>> [3,]    1024.5     908.8       227    1594
>>> [4,]   26267.0   27674.0     16199   22104
>>> [5,]     130.0     193.0       168     145
>>>
>>>
>>> I also notice that the dimension of the exprs matrix is such that
>>> there is one column for each chip, and as many rows as there are pm
>>> plus mm values.
>>>
>>> Are the first half of rows the pm values, with the mm values
>>> following, or are the pm values every other row with the
>>> corresponding
>>> mm value below, or is this set up in some other way? Is there any  
>>> way
>>> for me to look at a value in the exprs matrix and find out which
>>> entry
>>> in the pm/mm value list it is?
>>
>> The chip is read in row-wise, and the PM probes are in a given row,
>> with
>> the MM probes in the following row. Therefore, the data (excluding  
>> the
>> various QC probes) will be N PM probes followed by N MM probes,
>> where N
>> is the row length of the chip.
>
> This is not true I believe. The are no clear order of the pm and
> mm's. You need to get that information from somewhere else, usually
> from a CDF file.
>
> Karin: you will need to use the makecdfenv package to make what is
> called a CDF package - an R representation of the PM/MM/probeset  
> pairs.
>
> Kasper
>
>
>> If you really want to work with the exprs matrix directly (why?), you
>> can use indexProbes() to find the indices for whatever probeset you
>> are
>> interested in, and then subset out. Alternatively you can get the
>> indices for the PM and MM probes and subset those out separately
>> (which
>> is how pm() and mm() work). You can also use pm() or mm() with an
>> optional genenames argument to get the PM or MM probe values for a
>> particular probeset or probesets.
>>
>>
>> Best,
>>
>> Jim
>>
>>
>>>
>>> TIA,
>>>
>>> Karin
>>
>>
>> -- 
>> James W. MacDonald, M.S.
>> Biostatistician
>> Affymetrix and cDNA Microarray Core
>> University of Michigan Cancer Center
>> 1500 E. Medical Center Drive
>> 7410 CCGC
>> Ann Arbor MI 48109
>> 734-647-5623
>>
>>
>> **********************************************************
>> Electronic Mail is not secure, may not be read every day, and
>> should not be used for urgent or sensitive issues.
>>
>> _______________________________________________
>> 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