Correlation Network with Cytoscape

Question

[gina02]

I performed differential expression analysis of two different diseases and I have a list of DE human genes. I also have the logfc and p values .I want to visualise these in Cytoscape (as a correlation network between the 2 diseases). Are the logfc and pvalues enough for constructing a correlation network or do I need more? How should I choose which is the source node and the source attributes?

[Marco Pannone]

When I first used Cytoscape to perform network analysis based on my DEGs, I watched this video tutorial on YouTube and I found it very useful. I think for certain procedures is better to learn by watching a video than trying to explain in a comment. Hope it will help :)

Answer 1

[scooter]

Typically, all you really need to produce a correlation network is the Log2 fold change. The p-Values should be used to drop any edges that are not significant. Also, you don't need any edges at all (those are calculated as part of creating the correlation). There are a couple of apps that will help you create your network. The one I'm most familiar with is clusterMaker2, so you can give that a try, but there are certainly others you can look at.

-- scooter