[BioC] expresso: performing RMA on NON-Affy data?

[Kasper Daniel Hansen]

On Apr 23, 2009, at 1:27 , J.delasHeras at ed.ac.uk wrote:

> Quoting "James W. MacDonald" <jmacdon at med.umich.edu>:
>
>> Hi Jose,
>>
>> Do you want to do RMA, or just normalize? The problem with trying to
>> wedge things into an AffyBatch is that the affy package will then try
>> to find the cdfenv that contains the probe to probeset mappings, so  
>> by
>> trying to leverage the AffyBatch infrastructure you will have to also
>> come up with a fake cdf.
>>
>> If you don't have probes that make up a probeset, then I'm not sure  
>> the
>> affy package will be of use here.
>>
>> Can you give more details?
>>
>> Best,
>>
>> Jim
>
> Hi Jim,
>
> normalisation is not an issue, it's more to do with the  
> summarisation of probesets and something like 'Expresso' seems like  
> a good way to do what I need (and some other things I don't need).
>
> I am dealing with Nimblegen arrays. Both two colour (whole genome  
> promoter arrays, with anything up to 20 probes per probeset), and  
> one colour "a la Affymetrix" (expression arrays, with anything  
> between 3 to 8 probes per probeset).
>
> I've been dealing with teh two colour stuff just like I used to deal  
> with my spotted cDNA arrays, using Limma. To summarise the data...  
> I've used several approaches. Mostly I am not interested in the  
> whole 2.7kb that each "promoter region" comprises, so I've taken  
> subsets blah blah... Anyway, I'm happy with the results there.
> But for the expression data, I have one channel data, just like Affy  
> data. Numblegen provides already normalised and summarised data  
> along with the raw data, and they state they use the RMA procedure  
> which I've come across with when readingabout Affy chips, although  
> I've never analysed Affy data myself.
>
> I'm reasonably happy with the data given to me. It looks reasonable.  
> So I want to be able to do that myself rather than depending on  
> their data (thus allowing me to do things a bit differently if I  
> want to), and since the RMA-processed data looks good, I am  
> interested in finding a way to do RMA myself.
>
> You're right, the problem with my trying to make an AffyBatch from  
> non Affy data is that I'm going to have to create a cdf-like file...  
> and probably will encounter other obstacles... that's why I thought  
> I'd ask here, as there's people who are very familiar with that  
> structure...
>
> In my naivety, it seems it should be a simple enough task... and as  
> I'm using 4 types of arrays mostly... I'd only have to do some work  
> to make these work and then just enjoy the ride as new experiments  
> roll in...
> Am I naive? ;-)

It is pretty simple to do what you want, but "simple" is of course in  
the eye of the beholder - it depends on how familiar you are with the  
affy structures from a development point of view.

I am not familiar with Nimblegen, but that might be much easier, as in  
working out of the box.

Kasper


> I hope I clarified enough what I'm after.
>
> Jose
>
>
>
> -- 
> Dr. Jose I. de las Heras                      Email: J.delasHeras at ed.ac.uk
> The Wellcome Trust Centre for Cell Biology    Phone: +44 (0)131  
> 6513374
> Institute for Cell & Molecular Biology        Fax:   +44 (0)131  
> 6507360
> Swann Building, Mayfield Road
> University of Edinburgh
> Edinburgh EH9 3JR
> UK
>
> -- 
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
> _______________________________________________
> 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