[BioC] clusterStab using a correlation distance

[James W. MacDonald]

Hi Daniel,

Daniel Brewer wrote:
> Hello,
> 
> I am interested in using the package clusterStab and in particular the
> clusterComp function as a way to test the stability of the clustering of
> the dendrograms I have produced from microarray data.  The problem I
> have is that I would like to use a correlation distance (1-cor()) and it
> appears that the only option is to use a Euclidean distance.  Is that
> correct?

Yes, the only option is to use Euclidean distance.

> 
> Is there a way to alter the input data so that a Euclidean distance is
> in effect a correlation distance?

No, the distance matrix is computed internally, so inputting anything 
but a matrix of data will just make a mess of things.

> 
> If not would is be easy to change the source of the function?

Sure. Just download the source package, make whatever changes you want, 
and then build and install.

Best,

Jim


> 
> Many Thanks
> 
> Dan
> 
**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues