Application of Supervisory Control Theory to Guide Cellular Dynamics

Optimality in the control of gene regulatory networks

“…In previous works, we showed, first qualitatively [10] and later more rigorously [11], how one could model and control gene regulatory networks whose dynamics are amenable to a discrete-event representation. In these cases, we assumed that: a) genes have boolean states; and b) it is possible to "program" synthetic genes and insert them in the cell to be controlled, in a way that these genes realize the control actions prescribed by the mathematical formulation of the supervisor.…”

“…It is important to stress that the following rules are only a shortcut to obtain G hi . They are, in fact, the result of a procedure that was first laid out in [10] and later thoroughly and formally developed for the general case in [11]. It involves: a) enlarging the automaton G with the model of the synthetic gene layer; and, then, b) building an abstraction, based on state aggregation, to obtain G hi .…”

“…In previous works, we showed how ideas from Supervisory Control Theory (SCT) could be applied to solve the state attraction problem, which consists of guiding a cell from an initial state to some prespecified target state, keeping the system at this state indefinitely [10,11]. There, we assumed that: a) the plant consists of a gene regulatory network that can be abstracted by a discrete-event model; and b) the controller is implemented by synthetic genes that are inserted in the cell to close the loop.…”