How to validate existing networks with co-expression

Question

[loic.mangnier.1]

Hello there,

I currently work on regulatory network building based on Hi-C and I would like to validate them with a co-expression based approach. Briefly, I have clusters with several enhancers and promoters which are linked each other (the cluster size is moderate, between 2 and 20 elements in average) and I would like to show similar patterns in expression for the genes of the same cluster. I looked at the co-expression through expression correlation for different replicates, but I am not sure to integrate this to my existing networks. I wonder if simple methods could be used to illuminate the expression behaviour of my genes in the same cluster without proceed to resampling methods.

Any insights are welcome.

Thank you,

Loic

Answer 1

[Alex Nesmelov]

Check WGCNA functions moduleEigengenes and kME/corAndPvalue. From WGCNA manual it seems like kME is applicable only to signed gene networks, but I think that it will be applicable if patterns of all your GOI expression are positively correlated to each other - there shouldn't be a difference between signed and unsigned network in this case. I may miss some other restrictions, so please check WGCNA docs.

[loic.mangnier.1]

Thank you Alex for your answer. I will take a look to WGCNA module. Best