EnhancerNet: a model for enhancer selection in dense regulatory networks recapitulates the dynamics of cell type acquisition

Abstract

Multicellular organisms are made of many cell types that share a common genetic code. Cell type specification occurs through feedback interactions between transcription factors (TFs) and enhancers, which are regulatory elements with dynamic chromatin that can initiate transcription in distal genes. Understanding the structure and dynamics of these regulatory networks is a key challenge for cell biology. Here, we address this by considering a generic model for a dense feedback network involving TFs and enhancers, where TF binding modulates enhancer activity by adjusting its chromatin state ("EnhancerNet"). We show that experimental observations on autoregulation in these networks provide exacting constraints on the EnhancerNet model. The model recapitulates the dynamics of enhancer selection and cell type specification. EnhancerNet has two pathways for a cell to change its identity - by direct reprogramming or by differentiation through progenitors. These pathways capture the known experimental phenomenology of the corresponding pathways in cells. Moreover, they recapitulate known reprogramming recipes and the complex differentiation hierarchy in hematopoiesis with no fitting parameters. In addition to specific experimental predictions, the EnhancerNet model can explain how new cell types can evolve and highlights the functional importance of various aspects of gene regulation in animals.