What Are The Softwares Can Be Used For Hierarchical Clustering?

Question

[Priyanka]

Hi Biostars!

I have generated a matrix of numbers and wanted to do hierarchical clustering. Could anyone suggest me any tools or softwares for hierarchical clustering of the matrix (which is in csv format in a excel sheet).

Thanks

Answer 1

[seidel]

A really easy to use, general tool for clustering numbers is MeV (MultiExperiment Viewer), that originally came from TIGR and has been publicized by John Quackenbush for years. It's free, java-based, runs on any platform, has many tools for clustering and working with clusters, and is designed to be simple and easy to use. It's easy to install, and is the most straightforward way I know for most users (especially biologists) to cluster a table of data, and modify color themes and scaling.

Answer 2

[Obi Griffith]

When I first starting clustering data I found this page to be helpful. It includes links to both the original and improved versions of cluster/treeview. It also links to clustering libraries for python, perl and C. The latter was useful to me when I starting having memory/runtime issues with some of the other methods (if you have very large input files). However, these days I always go to R for clustering. The 'R & Bioconductor Manual' by Thomas Girke has a nice tutorial on clustering in R that covers all manner of clustering approaches with code examples and exercises.

[seidel]

The benevolence of Alok Saldanha's greeting: "Hello How are you Gentlepeople!" on the TreeView page is infamous, and will never wear off. I too use and recommend R for clustering (though I assumed that @Priyanka's csv format/excel comment implied no familiarity with R). That tutorial you pointed to is really nice.

[Priyanka]

Thanks @Obi... i tried using R commands but it showing a error "Error in if (n < 2) stop("must have n >= 2 objects to cluster") : argument is of length zero"

[seidel]

You have to call a distance function on your data matrix before giving it to hclust. For instance: hclust(dist(myData)). Look at the examples in the tutorial above.

[Priyanka]

Thanks a lot :-) nice tutorial i must say..

Answer 3

[Echo]

The cluster and treeview will help you a lot! you can have a try!

Answer 4

[Paulo Nuin]

Just use R

hclust

[Priyanka]

i tried using R. the error its showing is "Error in if (n < 2) stop("must have n >= 2 objects to cluster") : argument is of length zero"

[Madelaine Gogol]

If you post your code here, someone might be able to help you...