Hi - I am new to using bioinformatic tools like WGCNA, and I wanted some feedback on an intended approach for analyzing a dataset I am currently working on. I am working on using WGCNA to effectively analyze an RNAseq dataset that contains transcriptomes from 3 brain regions for 23 individual zebrafish (Danio rerio). The zebrafish belong to two lines that have been selectively bred for bold (n=11) and shy (n=12) behavioral types. I am working through tutorials on the website and other web resources. Following the details set out on pages 20-24 of this resource for analyzing multiregional data, I think the best approach is to use a consensus network to answer the following questions: A. In which brain regions is each consensus module expressed? B. What consensus modules are associated with fish from the bold or shy line?

My intended approach is as follows:

1. Create separate signed WGCNA networks for each brain region (3 total networks across bold/shy group) given that all individuals in this dataset belong to a single species.
2. Create a consensus network across brain regions and compare individual brain region networks to consensus network modules. I should be able to do this by creating an overlap table between individual region modules and consensus modules, as in this tutorial. This should answer question (A).
3. Correlate consensus network module eigengenes with traits (bold vs shy, male vs. female) to compare relative expression of modules associated with each trait. This will give me significance measurements that tell me whether a module is correlated with the bold or shy line, as in this tutorial. This should answer question (B).

I am intending to use this approach because it seems the most straightforward to interpret while still maintaining the multiregional nature of the dataset. Does this approach seem appropriate, or is there another approach that might model the multiregional nature of my data? Any advice is appreciated.