Evolutionary Robustness of Differentiation in Genetic Regulatory Networks

We investigate the ability of artificial Genetic Regulatory Networks (GRNs) to evolve differentiation. The proposed GRN model supports non-linear interaction between regulating factors, thereby facilitating the realization of complex regulatory logics. As a proof of concept we evolve GRNs of this kind to follow different pathways, producing two kinds of periodic dynamics in response to minimal differences in external input. Furthermore we find that successive increases in environmental pressure for differentiation, allowing a lineage to adapt gradually, compared to an immediate requirement for a switch between behaviors, yields better results on average. Apart from better success there is also less variability in performance, the latter indicating an increase in evolutionary robustness.