Drivers of structural features in gene regulatory networks: From biophysical constraints to biological function

• In gene regulatory networks (GRNs), network function drives network structure.
• Phenotypic constraints shape the edge usage and motifs frequencies in GRNs.
• In silico studies provide strong evidence for convergent evolution in GRNs.
• The impact of phenotypic constraints on GRN structure is highest for sparse networks.
• GRNs are sparse in models including the biophysical nature of interactions.

Living cells can maintain their internal states, react to changing environments, grow, differentiate, divide, etc. All these processes are tightly controlled by what can be called a regulatory program. The logic of the underlying control can sometimes be guessed at by examining the network of influences amongst genetic components. Some associated gene regulatory networks have been studied in prokaryotes and eukaryotes, unveiling various structural features ranging from broad distributions of out-degrees to recurrent “motifs”, that is small subgraphs having a specific pattern of interactions. To understand what factors may be driving such structuring, a number of groups have introduced frameworks to model the dynamics of gene regulatory networks. In that context, we review here such in silico approaches and show how selection for phenotypes, i.e., network function, can shape network structure.