You can use kmeans clustering which takes a number of clusters as a parameter. Also, you can use hierarchical clustering (see ?hclust) and cutree (see ?cutree) to cluster samples and then divide them into classes based on the clustering. Sean On Aug 12, 2004, at 8:20 AM, Liu, Xin wrote: > Hi, > > Such as in Golub data, he divide the data into four groups by setting > the group number as 4 in SOM. IF I use condition tree or principal > component analysis, I can not set the group number by myself. > > Xin > > -----Original Message----- > From: Sachin Mathur [mailto:smathur@kumc.edu] > Sent: 11 August 2004 16:46 > To: Liu, Xin > Subject: Re: [BioC] som - can it divide the samples into clusters > > > > Hi, > > You can cluster the samples in GS by using the condition tree and find > out the relations from the branch length or use principal component > analysis. > > Sahin. > >>>> "Liu, Xin" <xin.liu@arragen.com> 8/11/2004 3:22:05 AM >>> > Dear all, > > Somebody used a SOM to divide the samples into clusters. However the > SOM in GeneSpring only divides the genes into clusters, not samples. I > wonder the SOM in R can divide the samples? Or anybody knows a SOM can > do this work? Thanks a lot! > > Xin LIU > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor