[BioC] data transformation

Wolfgang Huber w.huber at dkfz-heidelberg.de
Wed Mar 17 21:15:19 MET 2004

Hi Rohit,

the results of vsn are invariant under scaling, i.e. the
result of


is the same as that of


except for an overall additive offset of the transformed values by a 
value of log(1000).

Since log-ratios (strictly speaking, "generalized log-ratios") are 
differences between transformed values, it makes no difference between 
the two calls.

Hope this helps,

Rohit Ghai wrote:
> hello all
> I am trying to perform VSN normalization with one-color Codelink chip data.
> after normalization in vsn i start getting negative values all over the 
> expression
> matrix... this makes it difficult to get fold changes for the genes. how 
> can i get
> around this problem ? I think its because the raw intensity values i get 
> from codelink
> are a bit low. I also multiplied the raw data matrix by a factor of 2. 
> With the trasnformed
> data matrix VSN did not give any negative values, but the boxplots 
> looked a little upward shifted
> as compared to the non-trasnformed data. The final gene list that I got 
> (genes with a standard
> deviation of more than 2 in any experiment) was the same both with the 
> raw data and with the
> transformed data.. but I want to know if its alright to do this?
> any ideas would be appreciated
> Rohit
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor

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

More information about the Bioconductor mailing list