[BioC] li.wong vs fit.li.wong

Rafael A. Irizarry rafa@jhu.edu
Thu, 26 Sep 2002 15:27:05 -0400 (EDT)


fit.li.wong
is ment to follow li and wong's paper.
takes an I x J matrix containing the probe set data. Typically
the i,j entry will contain the PM-MM (or PM) value for probe pair j
in chip i. 

li.wong is ment to work in the context of the express function. so its the 
transponse as you have noticed.

fit.li.wong takes in a matrix and returns the fitted phi's and thetas 
(among other things) of  the li and wong model. it doesnt normalize. the 
"invariantset" option in the 
normalize function does something similar to what dChip does. 

hope this helps
rafael

On Thu, 26 Sep 2002, Tibor van Rooij wrote:

> Hi! do li.wong( ) and fit.li.wong( ) have their matrix indeces reversed?
> 
> when I use the example:
> 
> x <- sweep(matrix(2^rnorm(600),30,20),1,seq(1,2,len=30),FUN="+")
> 
> and I use li.wong(x,verbose=T)
> 
> it claims it has :
> 
> chips used= 20 , probes used= 30
> 
> however when I use fit.li.wong(x,verbose=T)
> 
> it says it has:
> 
> chips used= 30 , probes used= 20
> 
> Does anybody know why? What I really would like to do is give it a 
matrix and get a matrix with of the same dimension with the normalized 
values back as I can with normalize.quantiles(x) for quantile 
normalization.   > 
> Has anybody succeeded in doing this?
> 
> Tibor
> _______________________________ 
> Tibor van Rooij 
> Bioinformatician 
> Montreal Genome Centre
> McGill University, Duff Building 
> Tel:  (514) 398-2497
> Fax: (514) 398-1838
> 
>