[BioC] vsn in BioConductor 1.2

[w.huber at dkfz-heidelberg.de]

Hi Charles,

vsn is a normalization method that brings the different columns (colors,
arrays) of an expression matrix on the same scale. As input, it takes an
n*d matrix, with d>=2. You passed it a matrix with d=1. Apparently this
results in some of the likelihood calculations becoming singular, hence
the error message you received.

Action plan:
1. For you: read the paper on vsn, then call it with expression matrices
of size d>=2.

2. For me: fix vsn so that it throws an intelligible error message if
called with d<=1.

Best regards
  Wolfgang

-------------------------------------
Wolfgang Huber
Division of Molecular Genome Analysis
German Cancer Research Center
Heidelberg, Germany
Phone: +49 6221 424709
Fax:   +49 6221 42524709
Http:  www.dkfz.de/mga/whuber
-------------------------------------


On Thu, 17 Jul 2003, White, Charles E WRAIR-Wash DC wrote:

> Could someone help me interpret (develop an action plan to correct ...) the
> error message that follows?
>
>
>
> Thanks.
>
>
>
> > Monkey.sub<-Monkey.expr[!is.na(Monkey.matrix[,1]),1]
>
> > Monkey.sub
>
> Expression Set (exprSet) with
>
>       13838 genes
>
>       1 samples
>
>              phenoData object with 6 variables and 1 cases
>
>        varLabels
>
>             : Slide
>
>             : FileName
>
>             : Cy3
>
>             : Cy5
>
>             : date
>
>             : Comments
>
> > Monkey.sub<-exprs(Monkey.expr[!is.na(Monkey.matrix[,1]),1])
>
> > Monkey.vsn<-vsn(Monkey.sub)
>
> vsn is working on a 13838 x 1 matrix, with lts.quantile=0.5; please wait for
> 11 dots:
>
> .Error in optim(par = p0, fn = ll, gr = grll, method = "L-BFGS-B", control =
> control,  :
>
>       L-BFGS-B needs finite values of fn
>
>
> 	[[alternative HTML version deleted]]
>