I have applied WGCNA on a data set which constitutes of samples with three phenotypes: control, diabetic and obese. There is one gene of interest say "X" for me which is expressed in all these conditions i.e. controls obese and diabetic. However, in obese and diabetic its expression is say 100/200/300 times increased.  I am only interested to construct a co-expression network from this data set in order to find a module which contains my gene of interest and perform downstream analysis like GO and Pathways etc. 

I feel no compulsion for doing module-trait analysis downstream of modules identification. Can anybody  suggest if it is rational approach?

I mean since I have already checked that X is a DEG in control versus obese and control versus diabetes in my data and i am only interested to find the genes highly co regulated with my gene of interest. so is it still necessary to perform module trait relationship or i can just continue with the module containing my gene of interest.