Co-option and Irreducibility in Regulatory Networks for Cellular Pattern Development

We used a computational approach to examine three questions at the intersection of developmental biology and evolution: 1) What is the space available for evolutionary exploration for genetic regulatory networks (GRNs) able to solve developmental patterning problems? 2) If different GRNs exist that can solve a particular pattern, are there differences between them that might lead to the selection of one over another? 3) What are the possibilities for co-opting GRN subcircuits or even entire GRNs evolved to solve one pattern for use in the solution of another pattern? We used a Monte Carlo strategy to search for simulated GRNs composed of nodes (proteins) and edges (regulatory interactions between proteins) capable of solving one of three striped cellular patterning problems. These GRNs were subjected to a knockout procedure akin to gene knock-outs in genetic research. Knockout was continued until all individual network components of the reduced GRN were shown to be essential for function. This GRN was termed irreducible. We found many different unique irreducible GRNs that were able to solve each patterning problem. Since any functional GRN must include an irreducible GRN as a core or subgraph, the space for evolutionary exploration of pattern-forming GRNs is large. Irreducible GRNs that solve a particular pattern differed widely in their robustness - the ability to solve a target pattern under different initial conditions. These differences may offer a target for selection to winnow out less robust GRNs from the set of GRNs found in nature. Finally, subgraph isomorphism analysis revealed great potential for co-option during evolution. Some irreducible GRNs appear in their entirety within larger GRNs that solve different patterning problems. At much higher frequency, subcycles are shared widely among irreducible GRNs, including those that solve different patterns. Irreducible GRNs may form the core elements of GRNs found in biological systems and provide insight int- o their evolution