Comparison of model predictions with experimental data.

<p>The discs mark the experimental mean, while the shaded region shows the standard error of the mean. Simulation results, denoted by the dashed line, show the total number of bacteria, <i>B</i>_<i>T</i>(<i>t</i>) (= <i>VB</i>_<i>F</i>(<i>t</i>) + <i>A</i>_<i>r</i><i>B</i>_<i>B</i>(<i>t</i>)). Note that the range of the <i>y</i>-axis for the untreated (control) scenario (left-hand column) is twice that for the single inhibitor dose scenario (right-hand column). There is good agreement between the model and the data in all cases and the overall goodness of fit (mean squared error) is similar for each parameter set. Eqs <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1006071#pcbi.1006071.e002" target="_blank">1</a>–<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1006071#pcbi.1006071.e016" target="_blank">8</a> were solved using ode15s and fitting was performed using a combination of MCMC and frequentist methods (see Parameter fitting and <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1006071#pcbi.1006071.s001" target="_blank">S1 Supporting Information</a>). See Tables <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1006071#pcbi.1006071.t001" target="_blank">1</a> and <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1006071#pcbi.1006071.t002" target="_blank">2</a> for parameter values. CFU: colony-forming units.