We recognize that there are many strategies available for analyzing complex datasets, such as those presented here, and each will emphasize somewhat different aspects of the data. The approach taken here is one that we have used previously (Colville et al., 2017; Iancu et al., 2018). The key metrics; DE, DV, and DW, are computationally straightforward and can be easily replicated. The WGCNA has greatly matured since its introduction (Zhang and Horvath, 2005) and has been used in more than 300 publications. In the current study we have focused our investigations on those genes that contribute to at least 80% of network connectivity. This thresholding reduced the number of genes considered for further analyses from ∼15,000 to ∼6,500 in each of the three brain regions. The genes culled are “leaf” nodes with low connectivity. While selection will have significant effects on some of these culled genes, none will be hub nodes. We also note that the sample sizes used in the current study were sufficient to produce networks of high quality (Langfelder and Horvath, 2008). The selection of the High
