[BioC] normalize.invariantset usage

[Laurent Gautier]

On Mon, Jan 05, 2004 at 03:13:32PM -0500, david neil hayes wrote:
> I am interested in normalizing an expression matrix (not an affybatch) 
> using the function "normalize.invariantset".  I have had luck with the 
> related normalization functions: normalize.constant,  normalize.loess, 
> normalize.quantiles (to name a few).
> 
> Normalize.invariantset, unlike the other functions mentioned, does not 
> accept as its input a matrix of expression values, but rather a vector of 
> expression values (presumably from one chip), as well as a vector of 
> expression values against which to normalize (presumably a second chip?).  
> The function returns a list of length 2 comprised of 1 object of class 
> "smooth.spline" and 1 object which is a logical vector of the genes 
> comprising the invariant set.

Actually the input is the data from two arrays (one of which being the
'reference' array against which the normalization is performed).
This was implemented this way for memory usage concerns (that are becoming 
less and less valid in time), and as you point it out it appears no so 
consistent with the other methods.

> 
> What I don't understand is how to obtain the normalized expression values 
> for all gene in the vector from the smooth.spline object (The smooth.spline 
> only seems to contain normalized expression for the invariant set).  I 
> suspect it is simple, but I have been unable to figure this out.

The following outlines what is done:

data(affybatch.example)
x <- exprs(affybatch.example)[,1]
ref <- exprs(affybatch.example)[,2]
tmp <- normalize.invariantset(x, ref)
x.norm <- as.numeric(approx(tmp$n.curve$y, tmp$n.curve$x, 
                  xout = x, rule = 2)$y)

(PMs and MMs are are handled differently in practice, you
you will refer to the code in normalize.AffyBatch.invariantset
for details)


Hoping it helps,


L.



> 
> Thanks in advance.
> Neil Hayes
> 
> using R1.8.1 on the Windows XP, package=affy1.3.26
> 
> _________________________________________________________________
> Tired of slow downloads? Compare online deals from your local high-speed 
> providers now.  https://broadband.msn.com
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor

-- 
--------------------------------------------------------------
Laurent Gautier			CBS, Building 208, DTU
PhD. Student			DK-2800 Lyngby,Denmark	
tel: +45 45 25 24 89		http://www.cbs.dtu.dk/laurent