Chemical Memory with Discrete Turing Patterns Appearing in the Glycolitic Reaction

abstract: Memory is an essential element of information processing devices. We investigated a network formed by just 3 interacting nodes representing continuously stirred tank reactors (CSTRs) where the glycolytic reaction proceeds as a potential realization of a chemical memory unit. Our study was based on the 2-variable computational model of the reaction. The model parameters were selected such that the system has a stable limit cycle and several distinct, discrete Turing patterns characterized by stationary concentrations at the nodes. In our interpretation, oscillations represent a blank memory unit, and Turing patterns code information. The considered memory can preserve information on one of 6 different symbols.  The time evolution of the nodes was individually controlled by the inflow of ATP. We demonstrated that information could be written with a simple and short perturbation of the inflow. The perturbation applies to one or two nodes only, and it is symbol specific. The memory can be erased with identical inflow perturbation applied to all nodes.  We believe the idea of pattern-coded memory applies to other reaction networks that allow for discrete Turing patterns. Moreover, it hints at the experimental realization of memory in a simple system with glycolytic reaction.