(617ck) Arrays of Coupled Cells Performing Chemical Computing with and without Parallelism Using Discrete Turing Patterns
We analyze two, three and four coupled cells in a linear and in a cyclic array. Each cell is modeled as a CSTR, coupling is assumed via membranes. Glycolysis is modeled using core model of glycolysis by Moran and Golbeter (1984), where ATP is transformed to ADP through phosphofructokinase and negative feedback is provided via lower parts of the glycolytic reaction chain whereby ADP is transformed back to ATP. We assume that rate coefficients of positive and negative feedback steps are modified via presence of other metabolites, pH and temperature, and thus may be elevated up to 9-fold the values at usual laboratory conditions. Transport coefficients are assumed equal for both ATP and ADP and this value is taken as a variable parameter.
Analysis of stability and bifurcations of stationary states provides parameter ranges corresponding to regions of coexistence of multiple discrete Turing patterns with oscillations, which is in turn used to examine transitions between discrete Turing patterns and oscillations using carefully targeted perturbations by ATP.
We propose a chemical computing technique capable of parallelism with a central knockout perturbation system, which replaces our previously published local knockout perturbation system. These techniques use transitions among discrete Turing patterns to achieve advanced logic functions and their plasticity during a chemical computing process.
In particular, we propose a chemical computing technique operating with a single data thread based on two coupled cells working as an XOR gate and using transitions among discrete Turing patterns and oscillations. This type of gate can be used to build a central processing unit.