Emma M. Myers¹, David Feng², Michael J. Hawrylycz³, Jason W. Bohland4
¹Boston University, ²Allen Institute for Brain Science, ³Allen Institute for Brain Science, 4Department of Health Sciences, Boston University
Large, anatomically linked gene expression datasets offer opportunities for comparing the brain’s molecular architecture across species. Here we used two such datasets to analyze correspondences in gene expression in homological areas of the adult human and mouse brain. Data were collated from the Allen Brain Atlases for mouse (based on anatomically registered ISH results) and human (based on microarrays performed on samples from six donor brains) using ~3000 orthologous genes. This study examined human-mouse brain homology at the molecular level in three ways. First, we compared gene expression profiles from throughout the brain in both species. Sample correlation matrices within and between species demonstrated clear regional organization, with blocks of higher correlations corresponding to coarse brain structure. Second, we compared homologous structures based on the expression of specific gene sets. Our goal was to determine if the expression profiles of different groups of genes were conserved in specific areas. We implemented a scoring system to assess the extent to which a given set of genes creates a region-specific molecular signature that is consistent across species, and identified gene sets that scored significantly higher than chance for several brain regions. Finally, each gene’s expression pattern across the mouse brain was correlated with its pattern across the human brain, and highly correlated genes were analyzed for overrepresented annotations. Overall, these results reveal substantial similarity in the molecular landscape of homologous brain areas in the mouse and human, but the strength of these relationships depends critically on the areas and gene sets examined.
Keywords: Gene Expression, Human, Mouse