[BioC] contrasting cluster images

[Benjamin Otto]

Here is a jpg-image demonstrating the effect of the z-scoring I mean. The
only difference in the three displayed cases is the application of the
z-score and when it is applied. The center image corresponds to point 2a in
my post and the right image corresponds to 2b. 

p.S.: I hope the image passes the admin curration, I couldn't get it smaller
than 160kb.

 Regards 
benjamin

-----Ursprüngliche Nachricht-----
Von: bioconductor-bounces at stat.math.ethz.ch
[mailto:bioconductor-bounces at stat.math.ethz.ch] Im Auftrag von Benjamin Otto
Gesendet: 03 November 2006 14:08
An: 'BioClist'
Betreff: [BioC] contrasting cluster images

Dear community,

 

When I look at Dchip cluster the expression value range is most of the time
between -2 and 2. That's because the standard options lead to a z-score
standardization of values.

 

1)       Does anybody know whether this procedure is performed before or
after clustering?

2)       The procedure seems to return nice visual contrasts for clusters of
already determined target genes. However it distorts the ratio/relationship
between the single genes AND samples

a.       visually if applied after clustering and

b.       in addition leads to different clusters if applied before
clustering.

So is there some good reason (beyond contrasting) to use the z-score BEFORE
clustering?

 

 

Regards

 

Benjamin

 

 

 

 

 

--
Benjamin Otto
Universitaetsklinikum Eppendorf Hamburg
Institut fuer Klinische Chemie
Martinistrasse 52
20246 Hamburg

 


	[[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