Home | Introduction | Documentation | Downloads | Mailing List & Contact | Links | References


2005 Dec 14

The latest version of CompClust 1.2 is now available.

Version 1.2 Includes a substantial revision of our principal component analysis (PCA) interpretation software (publication forthcoming).

Interested researchers can use our web-based application, CompClustWeb to interactively review the results of our analysis of the GNF human GNF human gene expression data from the SymAtlas web site at the Genomics Institute of the Novartis Research Foundation.

You can download one of the CompClust packages to futher explore CompClust. We have a new tutorial [ html | pdf ] that describes how to use the CompClust programming API to perform a PCA interpretation analysis of a dataset.

2005 Jun 22

Beta version of compclust 1.1 made available

Version 1.1 focuses on making installation from the python source much simpler and far more robust.

New features consist of shipping the clustering algorithm binaries, installing those binaries when doing python install and finding them after having installed compclust.

You can get the source If there are any problems please let us know. send email to user compclust with a domain

Upcoming plans include fixing rendering of the TrajectorySummary plot which gets crowded as the number of clusters increases. More work on trying to opensource the C code, switching to Quixote2, simpler data loading tools, and a mysterious new pca analysis tool.

2005 May 27

CompClust mailing list now available

If you're using CompClust please sign up so we can notify you of future releases and discuss what features to develop next.

2005 May 16

CompClust 1.0 release (previously known as 0.2.1)
CompClustTk 1.0 release (previously known as 0.2.16)
CompClustTk tutorial updated

In the new release, CompClustWeb, has made significant progress toward being able to duplicate the functionality of CompClustTk.

CompClustWeb as one might guess provides access to the core CompClust functionality from within a web browser. It can be used either from within an scgi compliant server or using one of the python webservers (like medusa or twisted).

CompClustWeb supports loading multiple datasets simultaneously, and has support for running the MultiRun cross validated clustering algorithm. CompClustTk was our first GUI toolkit and provides a significantly more interactive PCA browser, additionally accessing the IPython command line to perform tasks such as log2 transforming a dataset is much easier from within the TK version.

When run with one of the python webservers it is possible to access the datasets from within IPython although it is currently undocumented how one might go abou taking advantage of this.

Currently the CompClustTk version also has better documentation.

We hope to create a new release soon that will be the last supported version for CompClustTk that includes the ability to run MultiRun. We plan on making CompClustWeb our primary GUI environment in future releases.

Internally the methods for accessing wrapper parameters have gone through some significant changes and are hopefully much easier to use if one is writing CompClust scripts. (If you're interested the test cases contain examples of how these new features work).


  • Added information about CompClust, CompClustTk, and CompClustWeb to Introduction section.
  • Improved installation documentation.
  • Added binaries for Windows algorithms for those who wish to install the CompClust Python package. (Already included in CompClustTk Windows Installer.)


  • CompClustTk v0.2.12 Released
  • CompClust v0.2 Released

WoldlabWiki: compclust (last edited 2010-07-27 23:48:20 by diane)