2016 年 28 巻 2 号 p. 37-42
Sensitivity analysis is an efficient way to characterize network systems such as metabolic reaction and closed ecological-life-support systems. This analysis can be performed efficiently using logarithmic gains, rate-constant sensitivities, and kinetic-order sensitivities defined in Biochemical systems theory (BST). In the present study, the steady-state sensitivities calculated by three calculation methods based on BST, i.e., S-system, GMA-system, and modified GMA-system methods, are compared to elucidate the performances of their methods and then to determine the best method for conducting correct sensitivity calculation. Comparisons clearly show that the modified GMA-system method always provides correct rate-constant and kinetic-order sensitivities, whereas the S-system and GMA-system methods give partially incorrect values because a flux in only one of the differential equations is partially differentiated despite the fact that the same flux is contained in different equations. It is demonstrated that the summation of the relevant sensitivities with respect to the same parameter in different differential equations provides correct values.