[BioC] splitting clusters and cluster significance

Subramanian Karthikeyan Subramanian_Karthikeyan at hc-sc.gc.ca
Thu Sep 25 00:51:01 MEST 2003

HI Robert,

Thanks but I am not sure how cutree works.

I tried this:

x = read.delim("filename.txt", row.names = 1)
xdist = dist(x, method = "euclidean")
xclust = hclust(xclust, method = "complete")

this produces the nice dendrogram of clusters

cut1 = cutree(xclust, k = 2)

just produces a plot of (some) points in 2 D....no lines no cluster..what I
really wanted to see is a section of my original dendrogram so that I can
read the genes clustered together clearly.

how do we achieve this?

Thanks again.

Best regards,

                      Robert Gentleman                                                                                              
                      <rgentlem at jimmy.h        To:       Subramanian Karthikeyan <Subramanian_Karthikeyan at hc-sc.gc.ca>              
                      arvard.edu>              cc:       bioconductor at stat.math.ethz.ch                                             
                                               Subject:  Re: [BioC] splitting clusters and cluster significance                     
                      09/24/03 02:40 PM                                                                                             

On Wed, Sep 24, 2003 at 01:20:48PM -0400, Subramanian Karthikeyan wrote:
> Hi All:
> I am clustering ~500 genes using hclust of R.  Visualizing cluster
> membership in the dendrogram becomes impossible with so many genes in
> cluster. with all the labels overlapping..Is there a way of printing the
> dendrogram on multiple pages so that I can clearly see what is in each
> cluster? Or is there a way of splitting the tree so that I see what is in
> each cluster....?

  cutree  will cut the tree
> Again, how would I test the significance of clustering...i mean a measure
> of telling how tight my clusters are, and how separated they are from
> other.

  that's harder - there are books, chapters in books and many papers
  on the subject - Gordon's book, Classification (which is mainly
  about clustering, in the sense you are using the word) has some good
  explanations of different methods; but there are, as I said, many

> Thanks in advance.
> Karthi.
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor


| Robert Gentleman                 phone : (617) 632-5250
| Associate Professor              fax:   (617)  632-2444
| Department of Biostatistics      office: M1B20
| Harvard School of Public Health  email: rgentlem at jimmy.harvard.edu

More information about the Bioconductor mailing list