We are performing a co-expression analysis with WGCNA in addition to a standard differential expression analysis. We input our whole filtered normalized matrix into WGCNA (not DEG-filtered) and have identified modules highly correlated with our case trait. Within our top module, we have a list of all the module genes, and we would like to know if it is an acceptable practice to subset these module genes with the DEGs we have identified in our differential expression analysis, so that we have a list of the DEGs that are found in our top module. We want to use these DEGs in this module for generating our co-expression network. Is this an acceptable practice, or should we only use the determined hub genes based on the thresholds such as geneModuleMembership and geneTraitSignificance? Thank you.
