| Modeling in silico has become an extremely important tool in helping explain biological systems. To analyze properties and the behavior of complex cellular networks, methods that focus on systematic properties of the network are needed. Using only physical growth models is both expensive and extremely time consuming. In silico modeling can help reduce both time and energy exhausted in these studies. This work is to develop a mathematical model which will help optimize a strain for bioseparation using flux balance analysis. Flux balance analysis is used to create a system of linear equations which describe the rates for all metabolic pathways (approximately 740 reactions, 540 metabolites). By taking a flux balance over each metabolite, this set of equations takes the form S*x = b, where S is a -540 by 740 matrix, x is a column vector representing the reaction rates, and b is the flux of a given metabolite. This system is underdefined since there are more unknowns than equations; thus an optimal solution must be found. A routine was developed in MATLAB to choose an optimal root of this equation based on least squared differences. Once the model was developed, changes to a strain can be simulated by adding, removing, or altering equations in the flux balance analysis. Thus, the model should give a prediction of what should happen allowing only strains which are found by the model to be beneficial to be tested. The model was tested successfully against a battery of flux balance analysis models used on a smaller scale (approximately 30 reactions, 30 metabolites). It was then used to predict the ability of individual gene deletion mutants to grow. Future analysis would allow any combination of multiple gene deletion mutants to be tested with the ultimate goal being a bare bones genome which produces the fewest number of contaminants. |
