4.3.5. Gene Expression Network Correlation Analysis

The Allen Brain Institute provides expression data for about 25,000 genes, aligned in a 200 µm voxel-grid spanning the whole mouse brain [26,86]. A corresponding 3D mouse brain atlas (Common Coordinate Framework, CCF version 3 [25]) allows to link expression data to the respective brain regions. In our case, the more than 800 partially very small brain structures of the CCF were merged into greater regions to match our Paxinos reference atlas used for fMRI analysis (see Section 4.3.2). Mean expression data per gene were calculated by averaging all grid voxels belonging to the corresponding brain region.

As the used mouse models differ in their expression of sensory and sympathetic neurotransmitters, we tried to attribute the differences found in the brain networks to certain neurotransmitters. Therefore, we used the gene expression data provided by the Allen Brain Institute. For the neurotransmitters α-CGRP, SP and NA, the brain region-specific mean expression data of the typical receptors (target) and proteins such as transporters and synthesis enzymes (origin) were calculated for each neurotransmitter. In case that more than one gene was related to the origin (or target) of the transmitter, the maximum expression per brain region of all related genes was used. For visualization purposes, the logarithm to the basis 2 of the maximal expression values was used. The FC networks were then split up to represent only brain regions as nodes that express either the origin or target genes and the corresponding edges. Hereby, we aimed to access information about which neurotransmitters influence the changes in FC most. Results of this network representation were summarized in tables for easier access (Table 1, Table 2 and Table 3).