The coexpression network was constructed by means of the WGCNA (Langfelder and Horvath, 2008; Iancu et al., 2012). We started by constructing adjacency network matrices independently for each region by computing the Pearson correlation between all gene pairs. These values were raised to a power β = 6 for all regions, which was chosen such that the network approaches a scale-free structure (exponential distribution of node connectivity).
