ABOUT

MGclus is a new algorithm designed to detect modules with a strongly
interconnected neighborhood in large-scale biological interaction networks.
MGclus is implemented in Java and utilizes the JGraphT graph library. This 
file has information about how to use this tool.

INSTALLATION

The MGclus tool requires the jgrapht and log4j Java libraries. The MGclus 
package contains all required libraries and the source code.

The easiest way to use MGclus is to use the included JAR file. 

EXECUTION

The input is a tab-separated network file, with node names in the
first and second column, and an optional edge weight in the third
column. Via the command line interface, the user can specify the
merge-gain cut-off, whether edge weights should be used, and other
options.

Run the MGclus package as follows:

java -jar mgclus.jar -f INFILE [OPTIONS]

The output is a list of found clusters, each on a separate line as tab
separated node labels.

If not specified, the name of the output file is as follows:

mgclus.out.INFILE_NAME

EXAMPLE

There are two example input files included:

The first example contains a random network with 20 clusters. 

The second example contains S. Cerevisiae protein interaction data from BioGRID
(Stark et al., 2004). The network is restricted to high-throughput physical 
interaction data and was further limited to data from experiments utilizing the 
yeast two-hybrid technique.

References

Stark,C. et al. (2004) The BioGRID Interaction Database: 2011 update. 
Nucleic acids research, 39, 698-704