[BioC] cross validation / bootstrap after classification

Heike Pospisil pospisil at zbh.uni-hamburg.de
Fri Nov 4 17:02:00 CET 2005


> I have a package (ClassDiscovery) at
>     http://bioinformatics.mdanderson.org/Software/OOMPA
> that includes classes for
>     PerturbationClusterTest
> and
>     BootstrapClusterTest.
> Richard Simon's book (Design and Analysis of DNA Microarray Experiments) 
> includes a section on assesing the validity of clusters.

Thanks for these hints.

> Of course, clusters arising from a supervised selection of genes aren't 
> meaningful anyway....

I see, my explanation was too unexact. I use the t-test to get a sub set of gene 
and cluster them. Now, I would like to decide how robust is this selection 
depending on a random selection of chips.

Sorry for this confusion and thanks for your help.
Will try your package soon.

Best wishes,

More information about the Bioconductor mailing list