DATA ARE DOWNLOADED WITH jCLUST APPLICATION

Interface and intermediate files

The main interface looks like below. The users needs to load the file with the interactions and set the parameters of the algorithms like mentioned below.

This Panel shows the contents of the file that was loaded. In this case it shows the PPI interactions from the file that we loaded.

This Panel below shows the results of the clustering that were produced by the initial clustering methods.

This Panel below shows the intermediate results produced by the clustering methods.

This will be later the input for the secondary filtering methods methods. It contains information about the clusters formed and the interactions of the nodes within the clusters. these interactions were found in the initial input file format.

This file like every file is stored locally on the hard disk. This is because in the case of a big dataset, someone can load this file directly by skipping the time consuming run of MCL. This was done mainly for time efficiency purposes.

This panel which is the most important one shows the final results of the workflow after the clustering-filtering compunation.

Here we can see the number of clusters together with their elements. Results are also stored locally on the hard disk drive.

This is the help panel where someone can see the meaning of the various parameters of the algorithms.

This is the panel that holds information about Medusas input file. Medusa can be called within jClust application and the file can be loaded seperately. The option of visualizeing the predefined clusters will force Medusa to visualize distinct clusters. An example is given below.

Below we see how the initial interaction network look like and in bottom left image we see how layout algorithms help to isolate connections of specific nodes. In bottom right image we see how clusters look like. Of course Medusa comes with new interactivity and richer functionality which makes the exploration of these networks easier.