UCSF

As the brain develops, a vast diversity of neurons and glia are generated and distributed across the landscape of space and time. While in general this process is highly conserved across mammalian species, interspecies differences cascade into major changes in composition, scaling and ultimately the function of the adult brain. By generating single cell RNA sequencing (scRNAseq) data from across the period of rhesus monkey neurogenesis, we were able to characterize the distinct classes of inhibitory neurons in the cerebrum. With this taxonomy, we were able to identify instances of cell type evolution via generation of a novel primate-specific class of striatal interneurons, as well as differing distributions of olfactory bulb sister cell types between human, monkey and mouse. As cell type differences have their roots in developmental gene expression divergence, we next sought to define the mechanisms of gene expression evolution across the entire brain, which are not well characterized. To explore this, we developed a deep-learning based model of cell type evolution using scRNAseq to identify homologous cell types across species, call gene coexpression modules, and detect differential gene expression. We used this model to parcellate three forms of gene expression divergence and found that these changes were highly modular during both development and adulthood. We also show that the genomic context has a significant effect on whether a gene's expression will change in evolution. This work identifies multiple mechanisms of cell type evolution during embryonic development and moves toward a formalized model for how gene expression evolves in these cell types.