How do I select a subset of genes for functional enrichment(GO/KEGG) analysis from WGCNA results?

Question

I am working on using WGCNA for a bulk RNA-seq experiment. I have three experimental conditions (control, treatment 1, and treatment 2) and have a total sample size of 18 (6 per group). Using the online WGCNA tutorials, I could analyze my datasets after variance stabilizing transformation (vst). I got the summary of network analysis results as a CSV with Gene significance (GS), p.GS, module membership(MM), and associated p-value (p.MM).

I want to extract a set of significant genes from the module with the highest positive correlation to my treatment conditions. As I am just getting acquainted with WGCNA interpretation, which of these parameters in the CSV should I use for filtering further (GS, p.GS, and module membership and associated p-value (p.MM))?

Also, how would I choose the limits of the criteria? Should it be using the scatter plot (MM vs GS)?

As I am pretty new to WGCNA, any help is appreciated. Thanks in advance!

Answer 1

Consider filtering on every parameter: Gene Significance (GS), p-value for Gene Significance (p.GS), Module Membership (MM), and p-value for Module Membership (p.MM).

This doesn't mean you need to necessarily filter on every parameter, but it depends on what exactly you want to do, and you might reasonably choose any/all.

1. For GS, you may want to ensure you're capturing genes strongly associated with your experimental conditions.

2. For p.GS, you may want to ensure statistical relevance, (so filter out genes with high p-values). Similar for p.MM.

3. For MM, you may want it to be high to make sure the genes you select are strongly co-expressed and are relevant members of the module.

Plot MM vs GS to see if you can see an obvious filter threshold.