[BioC] How to include chromosomal location or GO-annotation data in supervised microarray analysis?

Wolfgang Huber huber at ebi.ac.uk
Thu Nov 6 20:52:11 CET 2008


Dear Rainer,

have a look at Claudio Lottaz's stam package:
Description: The stam package performs a biologically structured
         classification of microarray profiles according to clinical
         phenotypes. GO terms are used to link classification results
         to biological aspects. We call biologically focused signatures
         corresponding to these class predictions molecular symptoms.
         Thus, stam allows for molecular stratification of patients
         with complex phenotypes according to presence/absence patterns
         of molecular symptoms.

Best wishes
	Wolfgang



-- 
----------------------------------------------------
Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber

Sean Davis ha scritto:
> On Thu, Nov 6, 2008 at 1:56 PM, Rainer Tischler <rainer_t62 at yahoo.de> wrote:
> 
>> Hi,
>>
>> I have a microarray data set with additional information on the chromosomal
>> location of genes and their GO-groups. I'm looking for a simple way to
>> include this annotation data in a supervised microarray analysis (disease
>> outcome classification) to improve the prediction accuracy. There appear to
>> be two basic strategies:
>>
>> 1. combine similar genes to gene groups based on the annotation data before
>> starting the statistical analysis
>> 2. improve the distance measure for feature selection and classification by
>> including distance information derived from the annotation data
>>
>> Is anybody aware of an R-package that implements one of these ideas or is
>> there a simply way I could implement this myself (e.g. replacing gene groups
>> by a single gene based on the mean or median expression levels - I'm not
>> sure whether this would be effective or whether more sophisticated methods
>> are already available as R-packages)?
>> Currently, I'm using an SVM- and a PAM-classifier for my predictions, thus,
>> I hope to find an integrative approach which is compatible with these
>> classifiers.
>>
> 
> Check PGSEA or globaltest; there are likely others.
> 
> Sean
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> 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



More information about the Bioconductor mailing list