A Dynamic Cybernetic Approach: Optimal Control for Predicting Regulatory Metabolism Actions

In this work, a dynamic cybernetic modeling framework is introduced for in silico experimentation in the absence of partial kinetic information. In this method, cybernetic principles are employed, assuming that the biological system is evolved such that it optimizes a metabolic objective function. Based on this objective function, the missing dynamic information is evaluated through a dynamic optimization scheme. Log-linear models are used as the basis for cybernetic modeling, which enable obtaining analytical solutions to the problem. The existence of an analytical solution eliminates the computation time and globality related problems in dynamic optimization, and renders the proposed method well scalable to large problems. The ease in calculations provided by the analytical solutions renders practical in silico experimentation possible. These experiments can be used for rapid construction and testing of new hypotheses about kinetic expressions or checking the reliability of existing kinetic models. Further, it is demonstrated that the solution of the dynamic optimization problem can be used for inferring the entire regulatory mechanism and formulating log-linear models for the unknown kinetic rate equations. A glycolytic pathway example is studied for demonstrating the potential of the suggested approach.