Designing modular reaction-diffusion programs for complex pattern formation

Abstract

Cells use sophisticated, multiscale spatial patterns of chemical instructions to control cell fate and tissue growth. While some types of synthetic pattern formation have been well studied1-6, it remains unclear how to design chemical processes that can reproducibly create similar spatial patterns. Here we describe a scalable approach for the design of processes that generate such patterns, which can be implemented using synthetic DNA reaction-diffusion networks7,8. In our method, black-box modules are connected together into integrated programs for arbitrarily complex pattern formation. These programs can respond to input stimuli, process information, and ultimately produce stable output patterns that differ in size and concentration from their inputs. To build these programs, we break a target pattern into a set of patterning subtasks, design modules to perform these subtasks independently, and combine the modules into networks. We demonstrate in simulation how programs designed with our methodology can generate complex patterns, including a French flag and a stick figure.