The DEA Analysis In Metabolic Reaction Networks

Successful biotechnological applications, such as amino acid production, have demonstrated significant improvement in bioprocess performance by genetic modifications of metabolic control architectures and enzyme expression levels. however, the stoichiometric complexity of metabolic pathways, along with their strongly nonlinear nature and regulatory coupling, necessitates the use of structured kinetic models to direct experimental applications and aid in quantitative understanding of cellular bioprocesses. The S-system representation developed within biochemical systems theory (BST) allows the description of biochemical systems by nonlinear models of a power-law form.There are three main features that make this modeling approach attractive. Introducing the constraints containing binary variables in S-system model, Vassily Hatzimanikatiset al. Formulated the optimization problem as a mixed-integer linear programming (MILP) problem.These newly introduced constraints contain changes of a variety of enzyme regulatory architectures, reducing the amount of calculations. However, when reaction pathways become more complex, we need to solve the MILP model multiple times. And the model becomes very complicated as a result of introducing the new constraints. It is very difficult to solve the problem using linear programming method, and the optimal production rate can not be predicted as well. In order to get the optimal network architecture, this article predicts the production rate under different enzyme regular structures through utilizing the Data Envelopment Analysis (DEA) and objective programming. It is based on the result of resolving the MILP twice and optimization of the metabolic network is determined afterwards. In this paper, the development of Data Envelopment Analysis (DEA) theories, DEA methods and DEA applications are discussed. The basic DEA model, DEA efficiency theories are analyzed. we also analyze the solution of MILP framework, and then applied a DEA model to evaluate the relative efficiency of the solution. Furthermore, using projection, efficient DMU for DEA is constructed by the inefficient DMU. In this way, we can achieve the optimum regulatory architectures in metabolic reaction networks and improve the final product concentration.