Inference method for developing mathematical models of cell signaling pathways using proteomic datasets

The progress in proteomics technologies has led to a rapid accumulation of large-scale proteomic datasets in recent years, which provides an unprecedented opportunity and valuable resources to understand how living organisms perform necessary functions at systems levels. This work presents a computational method for designing mathematical models based on proteomic datasets. Using the mitogen-activated protein (MAP) kinase pathway as the test system, we first develop a mathematical model including the cytosolic and nuclear subsystems. A key step of modeling is to apply a genetic algorithm to infer unknown model parameters. Then the robustness property of mathematical models is used as a criterion to select appropriate rate constants from the estimated candidates. Moreover, quantitative information such as the absolute protein concentrations is used to further refine the mathematical model. The successful application of this inference method to the MAP kinase pathway suggests that it is a useful and powerful approach for developing accurate mathematical models to gain important insights into the regulatory mechanisms of cell signaling pathways.