A Gene Regulatory Model of Cortical Neurogenesis

Abstract Sparse data describing mouse cortical neurogenesis were used to derive a model gene regulatory network (GRN) that is then able to control the quantitative cellular dynamics of the observed neurogenesis. Derivation of the network begins by estimating from the biological data a set of cell states and transition probabilities necessary to explain neurogenesis. We show that the stochastic transition between states can be implemented by the dynamics of a GRN comprising only 36 abstract genes. Finally, we demonstrate using detailed physical simulations of cell mitosis, and differentiation that this GRN is able to steer a population of neuroepithelial precursors through mitotic expansion and differentiation to form the quantitatively correct complex multicellular architectures of mouse cortical areas 3 and 6. We find that the same GRN is able to generate both areas though modulation of only one gene, suggesting that arealization of the cortical sheet may require only simple improvisations on a fundamental gene network. We conclude that even sparse phenotypic and cell lineage data can be used to infer fundamental properties of neurogenesis and its organization. 1. Highlights Estimation of the cell states and transition probabilities of neurogenesis from experimental data. Design of an abstract gene regulatory network (GRN) whose dynamics implement cell states and their stochastic transitions. Detailed simulation of GRN-guided neurogenesis for mouse cortical areas 3 and 6. Different dynamics of neurogenesis of distinct cortical areas arise through modulation of only a single gene. 2. In brief Pfister et al. show how sparse phenotypic and cell lineage data can be used to infer a small abstract gene regulatory network (GRN), which, when inserted into model precursor cells, is able to control in a distributed manner the quantitative cellular dynamics of neocortical neurogenesis.