Hi,

Hopefully this should be quite an easy question to answer. I've produced consensus eigengene modules and network modules according to the instructions on the UCLA site. The earlier part of my study was looking at differential gene expression between 3 different conditions and I'm quite keen to see whether there is an over representation of DEGs within the different networks that I found. I'm new to WGCNA analysis and bioinformatics in general, so a way of looking at this problem didn't occur to me immediately.

Has anyone else done this and got any advice on how I might go about doing it?

Thanks!

The easiest thing to do is to correlate module eigengenes to the 3 different conditions as illustrated in WGCNA tutorial II, section 4. This will identify modules that relate to the condition.

What you wrote is a bit more roundabout, but you can also calculate the enrichment of the modules in DE genes. I don't have particularly elegant code handy, so here's a quick and dirty solution: create a vector that has one component per gene and value 0 if the gene is not a DE gene, and +/- 1 if it is negatively/positively differentially expressed. Call this vector deIndictator

Then run the function overlapTable(deIndicator, moduleLabels); the component pTable of the output will give you a table of enrichment p-values.

HTH,

Peter