Figure 3

Sensitivity analysis using the eFAST global sensitivity method on the thermodynamic model of Fakhouri et al., scheme 2. A) Shown are first- and total-order sensitivity indices, which represent the amount of variation in the model output with respect to each parameter individually (gray bars) and in conjunction with all other parameters (black bars), scaled from 0.0 to 1.0. If the gray bar is much smaller than the black bar, as in Q6, secondary (or higher) effects are predominant. Quenching parameters are generally most sensitive and cooperativity parameters are least sensitive. B) Effect of frequency with which a quenching parameter is represented in the twelve constructs on sensitivity. The number of constructs that the parameter is represented in is shown along the horizontal axis. Corresponding first- and total-order sensitivity indices are shown for each quenching parameter, as calculated by the eFAST algorithm. There are two quenching parameters (Q4 and Q5) represented in 2 constructs, and two quenching parameters (Q1 and Q3) represented in 5 constructs. At these values, there are four data points, two for each quenching parameter. The lines illustrate linear fits to each data set, first- and total-order sensitivity indices. In general, the more constructs a quenching parameter is represented in, the more sensitive that parameter.