Building & Comparing Gene Networks Based On Expression Values

Question

[Dataminer]

Hi!

I am kind of confused and lost with network building in cytoscape and the whole methodology.

What I have got.

I have a set of transcription factors for which a ChIP-Seq experiment was performed and annotating the peaks of these TFs with nearby genes I have a target geneset for each of the transcription factor. Ofcourse of the target genes are shared between the TF's.

What I did

I made the TF-gene network for this set of data using cytoscape.

The table looks like this

TF1 gene1 
TF1 gene2
TF1 gene3
TF2 gene2
TF2 gene4
TF2 gene5
TF3 gene1
TF3 gene2
TF3 gene3
TF3 gene4
TF3 gene6

What I want to integrate this network with expression values in normal and in diseased state and to compare the two networks in terms of their hubs and subnetworks and to identify network drivers for the two.

hence the complete set will look like this

TF1 gene1 1.3 2
TF1 gene2 0.9 2.3
TF1 gene3 2   0.3
TF2 gene2 0.9 2.3
TF2 gene4 4   2
TF2 gene5 6   3
TF3 gene1 1.3 2
TF3 gene2 0.9 2.3
TF3 gene3 2   0.3
TF3 gene4 4   2
TF3 gene6 0.8 3

column3 is for normal state and column 4 is for diseased state.

Thank you

[ancient_learner]

I think you have to build 2 networks(Normal Diseased) individually and check for the sub modules.

[Dataminer]

Agree, I think I just figured out what to do..... but will wait for the responses from others......may be I get more creative ideas :) Thank you

Answer 1

[Leandro Lima]

I recommend you to use Cytoscape filter to select the edges you want and create different networks.

After this, you can use Network Anlyzer plugin.

Answer 2

[Heather Vincent]

NetworkAnalyzer will find hubs. For finding subnetworks, have a look at the jActiveModules plugin.