Fig. 1

Strategy for identifying biologically relevant modules via network clustering analysis. Total RNA was extracted from the PAG and cortex of 6 adult bats. Sequencing was performed via Illumina paired-end RNA-Seq, followed by de novo transcriptome assembly. Following gene annotation and transcript quantification, it was possible to identify genes that were differentially expressed between the cortex and PAG. The major isoform of each protein coding gene was then selected to perform cluster analysis. Co-expressed genes were identified using two contrasting analysis models simultaneously on the same dataset; MCLUST and WGCNA. For both methods, to ensure high stringency, iterative re-clustering was performed to remove genes that were not strongly associated with the co-expressed modules. For each brain region, the modules identified via the contrasting techniques were compared in a pairwise fashion, and the most highly conserved networks were identified. Furthermore, these highly conserved networks were then compared across regions. This allowed us to identify a set of high confidence co-expressed transcriptional networks related to the PAG