Differential Coexpression Analysis

Question

[hkarakurt]

Hello everyone,

I want to do a differential coexpression analysis to a microarray data. My data have rows for genes and columns for conditions. 30 control and 28 disease columns.

I can build coexpression network for control and disease but I cannot find differentially coexpressed genes.

Also, I checked some packages such as WGCNA but I think they don't take expression data as input.

Anyone has any advice, tutorial, software or another package to do it?

Thanks.

[Arup Ghosh]

Do you want to build a co-expression network? https://github.com/arupgsh/2016-summer-workshop/tree/master/3B-Hughitt-RNASeq-Coex-Network-Analysis

[Kevin Blighe]

I think that the idea is to build two separate co-expression networks, one for cases and the other for controls, and then compare them. This is called differential co-expression analysis.

As far as I know, there are no methods that can do this for you, so, a manual curation of the differences via metrics such as vertex degree, BC and CC (betweenness and closeness centrality), and hub score, is the way to go. You can also look at differences in community structure between the cases and control networks.