Fig. 3

Effectiveness of selected evolutionary algorithms for estimating parameters of generalized mass action model under simulated conditions of increasing measurement errors. A Quality of estimations at increasing measurement errors as quantified by R²-value. Each of the three open ovals represents a value derived from a distinct triplicate dataset for both metabolites and enzymes and a random seed solution, whereas a filled oval denotes the R²-value of their mean parameter values. The trend of the filled ovals is shown as a straight line. B Correlation plots of estimated parameter values (mean from the three initial seed solutions) against actual ones at increasing measurement errors for each algorithm. C Simulated metabolite dynamics based on the estimated parameters. In this case, the errors are sampled from a normal distribution centered on zero and a standard deviation equivalent to 7.5% of the underlying metabolite or enzyme concentration. Refer to Additional file 6: Data S5 for datapoints presented in the figure