Cytoscape ClusterMaker2. Use Expression Matrix, Network needed?

Question

[Benni]

I want to use the Cytoscape App ClusterMaker2 on data from an expression matrix (rows =genes, columns = conditiones). But I think I need an Network to use ClusterMaker2. How can I create a Network out of my matrix? Or is there a way to use ClusterMaker2 on an imported table?

Thanks

UPDATE: I found really good way to create a Network of my matrix: "you can just read your expression matrix in as a network. Go to File->Import->Network->File... and select the column containing your gene names as the Source column (green circle). Then change all of the other columns to be Source Node Attributes (right-click on the column to bring up the dialog to change the column type). Then import it. You now have a "network" with no edges. You can now create coexpression networks from that, or use the hierarchical clustering algorithms." https://groups.google.com/forum/?fromgroups=#!topic/cytoscape-helpdesk/7l0UlinH1K8

Answer 1

[Jean-Karim Heriche]

You can transform your data matrix into a graph adjacency matrix by computing a measure of similarity between the genes (if you want the genes to be nodes of the graph). The similarity values are the edge weights of the graph. To import into cytoscape, save the matrix in a tab-delimited text file with a .mat extension and use the aMatReader app.