Hello, I tried to cluster data from Affy U133A by normalising with gcrma then zscoring but I am get different values and results from when using the Eisen Cluster 3.0 software and other commercial software. I am wondering what is the best way to baseline transform the data after normalisation show the most variable data in the set that can be used to show the relationship in clustering. In Genespring they use baseling transformation as follows : Baseline to median of all samples: For each probe the median of the log summarized values from all the samples is calculated and subtracted from each of the samples. In Cluster 3.0 It is recommended to log transform the data and mean or median centre the genes to transform the data. What is the best way to go about base transforming (e.g. scaling, mean centering) the data in biocondcutor before clustering them? Thanks Ruppert.

On Tue, Sep 16, 2008 at 7:42 AM, Ruppert Valentino <ruppert7 at="" hotmail.com=""> wrote: > > Hello, > > I tried to cluster data from Affy U133A by normalising with gcrma then zscoring but I am get different values and results from when using the Eisen Cluster 3.0 software and other commercial software. I am wondering what is the best way to baseline transform the data after normalisation show the most variable data in the set that can be used to show the relationship in clustering. > There is not a "best way". And, of course, you will potentially get different clustering depending on the parameters that you use in the different software packages. The nice thing about R is that you can use pretty much any transformation you like combined with any clustering method and distance measure. As for the most variable genes, you will want to determine those BEFORE you Z-score transform. There are many messages in the list archives dealing with filtering genes. Hope that helps, Sean > In Genespring they use baseling transformation as follows : > > Baseline to median of all samples: For each probe the median of the log summarized values from all the samples is calculated and subtracted from each of the samples. > > In Cluster 3.0 > > It is recommended to log transform the data and mean or median centre the genes to transform the data. > > > What is the best way to go about base transforming (e.g. scaling, mean centering) the data in biocondcutor before clustering them? > > Thanks > > > Ruppert. > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >