Home

jClust is an easy-to-use standalone application for clustering analysis. It comes with several supervised and unsupervised algorithms to cluster and group heterogeneous data. jClust runs under any OS. The  calculations take place locally at the users machine. Java 6 or higher is required for the application to run.

 

The application for 32bit OS can be downloaded from

HERE

For 64 bit Systems just recompile the executable files

 

The algorithms supported by jClust application are:

 

	k-Means
	Affinity Propagation
	Spectral Clustering
	Markov Clustering (MCL)
	Restricted Neighborhood Search Cluster (RNSC)
	MULIC

 

For detecting cliques:

Bron-Kerbosch algorithm is supported

 

To filter and reduce the noise of the clusters 4 filters are supported. These are:

	Haircut operation
	Density Control
	Outside-Inside Methodology
	Best Neighbor

 

To visualize clusters:

we implemented a newer version of Medusa Visualization tool.

 

The main idea behind jClust project is to provide a strong collection of clustering algorithms that can be applied to various data to address different problems. Ideas of how this software can be useful for biologists for:

 

Abstract clustering of literature that are related between each other

	Microarray clustering
	Identification of protein families
	Prediction of protein complexes from protein-protein interaction data
	Prediction and visualization of homologous proteins
	Clustering of heterogeneous data to see connections between clusters
	Chemical clustering using Tanimoto distances
	....and many many other case studies