[BioC] reducing gene number in affybatch objects

[Laurent Gautier]

On Thu, May 22, 2003 at 05:07:46PM -0400, Rafael A. Irizarry wrote:
> you are using hgu95av2. so change it to the following:
> 
> 1) data(hgu95av2cdf) #probably already loaded... 
> 2) takeout <- ls(hgu95av2cdf)[1001:12626]
> 3) rm(list=takeout,envir=hgu95av2cdf)
> 
> notice that hgu95av2cdf is included in the affy package (at least up to 
> 1.1.2) so you dont need to load the library.

I would like to stress that hgu95av2cdf *WAS* included in the package.
It is now in a external package of example data (package 'affydata').

I would also like to comment on the approach you suggest. It is a nice
idea, but can shoot in one's foot. The normalization methods work
on "probes" (the slot 'exprs' in the AffyBatch). Some of these methods
work with PMs, some with all the probes without considering the mapping
(PM, MM or else). By modifying the associated cdfenv, one only modifies
the mapping. Comparison of the processing methods of the data could be
risky (and I would suspect that a teaching session would make students
have a glance at that). The use of a subset of affyids suggested before
is safe.

Finally, to answer the qustion about VSN and affy 1.0. It won't work out
of the box, but the mecanism for additional processing methods was already
there. You could port vsn for Plobs.



Hopin'  it helps,



L.



 
> 
> Hinnerk Boriss wrote:
> 
> > 
> > > i'd like to add there is a simple way to remove genes. its 
> > > not the best
> > > way but its simple and itll do the trick:
> > > say ab is your affybatch and hgu95acdf is your cdf before 
> > > your class type:
> > > 
> > > 1) library(hgu95acdf) #probably already loaded
> > > 2) takeout <- ls(hgu95acdf)[1001:12626]
> > > 3) rm(list=takeout,envir=hgu95acdf)
> > > 
> > 
> > Rafael, I seem to be missing something. I've done the following:
> > 
> > library(hgu95acdf) #probably already loaded
> > takeout <- ls(hgu95acdf)[1001:12626]
> > rm(list=takeout,envir=hgu95acdf)
> > > junkab <- ReadAffy(widget=T) # reading in 4 hgu95av2 chips
> > > junkab
> > AffyBatch object
> > size of arrays=640x640 features (12804 kb)
> > cdf=HG_U95Av2 (12625 affyids)
> > number of samples=4
> > number of genes=12625
> > annotation=hgu95av2
> > notes= 
> > 
> > Hmm, this still gives me all genes, even though the cdfenv has now only
> > 1000 genes. Am I doing something wrong here?
> > 
> > > AffyBatch's are not a simple as Plob's, but they make other things 
> > > more efficient. +
> > 
> > I know, and I do appreciate those other aspects!
> > 
> > > > > Yours dreaming of good old plob objects,
> > > > 
> > > > The release 1.0 of the package is still available for 
> > > download I think...
> > 
> > Laurent, will the release 1.0 be compatible with Wolfgang Huber's vsn
> > package? I am afraid not. 
> > 
> > Best,
> > Hinnerk
> > 
> > 
> 

-- 
--------------------------------------------------------------
currently at the National Yang-Ming University in Taipei, Taiwan
--------------------------------------------------------------
Laurent Gautier			CBS, Building 208, DTU
PhD. Student			DK-2800 Lyngby,Denmark	
tel: +45 45 25 24 89		http://www.cbs.dtu.dk/laurent