[BioC] singlearrayerrormodel

w.huber at dkfz-heidelberg.de w.huber at dkfz-heidelberg.de
Tue Nov 18 16:39:37 MET 2003

Hi Yeon-Ki,

I am not quite sure what sure what you mean by "single array error model"
but I suppose you mean error models for the individual intensities on a
spotted cDNA array rather than for their ratios or log-ratios.

There are some methods in the limma package for doing single-channel
normalization, I am not sure how much that is based on error models
(Gordon or Nathalie Thorne may be better in explaing this).

The package "vsn" allows to fit variance-stabilizing transformations,
which are closely related to error models. The error model used is the
two-component error model, i.e. an additive (constant) and a
multiplicative (proportional) error component.

Best wishes

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/abt0840/whuber

On Tue, 18 Nov 2003 kim750 at ggbio.com wrote:

> Hi! everyone,
> My name is Yeon-Ki Kim.

> Recently, single array error model seems to be prevalent in analyzing a
> microarray data.  Does anyone give me some suggestions whether I can do
> this on Bioconductor or if so, how I can do it?

> Thanks.
> Yeon-Ki

Yeon-Ki Kim, Ph.D.

More information about the Bioconductor mailing list