Comparison of possible models describing formaldehyde-tolerance phenotype transition processes of M. extorquens populations.

For each combination of culture condition and growth substrate, we used a stepwise procedure to evaluate nested models using a likelihood ratio test. Shown below are the best-fit values for the four fitted parameters in each of those models, and test results for each model. α: dependence of death rate on formaldehyde (h-1*mM-1). v: advection rate (mM*h-1). D: diffusion constant (mM2*h-1). (for v and D, mM denotes tolerance). b: dependence of death rate on individual tolerance level (mM tolerance / mM formaldehyde in medium). For the likelihood ratio test, name of the model used as the null model, as well as the χ2 value and p-value, are given. Gray shading: the best-supported model for that experimental scenario. Pseudo-R2 values for those models were: for formaldehyde selection, 0.973; for methanol regrowth, 0.993; for succinate regrowth, 0.991.