2.2 | Sensitivity analysis
We used the Morris Method to examine the sensitivity of different
objectives to various model parameters (e.g., enzyme concentrations or
kinetic constants; (Morris, 1991)). This method calculates a set of
derivatives (i.e., the change in an objective with respect to a
specified parameter) by sampling randomly within a range of parameter
values. Briefly, we used the SAFE toolbox with a radial method and Latin
hypercube sampling (r = 1,000 trajectories), and verified the
convergence of the elementary effects by calculating the first
derivative of each parameter’s normalized elementary effects with
respect to the number of model evaluations (i.e.,
|dEE/dn| < 0.0002) before reporting them
(Pianosi et al., 2015). We examined kinetic parameters associated with
methyl ketone biosynthesis (Fig. S6) by using parameter ranges of
0.01-100 s-1, and we explored the sensitivity of
different biochemical objectives to enzyme concentrations (Fig. 4) by
using 0.01-100 μM for all enzymes except those in the alkane pathway;
for this low-producing pathway, the narrower range of 1-100 μM enzyme
facilitated convergence.