10.1371/journal.pone.0215906.g017 Dustin F. Kapraun Dustin F. Kapraun John F. Wambaugh John F. Wambaugh R. Woodrow Setzer R. Woodrow Setzer Richard S. Judson Richard S. Judson Fetal kidney mass vs. gestational age. Public Library of Science 2019 time-varying quantities blood flow rates PBPK Akaike information criterion Environmental Protection Agency model quantity 2019-05-02 17:40:30 Figure https://plos.figshare.com/articles/figure/Fetal_kidney_mass_vs_gestational_age_/8070599 <p>The power law model (solid line) given by <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.e204" target="_blank">Eq 34</a> was selected as the most parsimonious model in our analysis; however, that model is a function of fetal mass. The Gompertz model (dashed line) given by <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.e205" target="_blank">Eq 35</a> may be preferred since it is a function of gestational age and does not require an intermediate model for fetal mass. Note that the Gompertz model and the power law model are virtually indistinguishable in this plot. The model of Luecke et al. [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.ref015" target="_blank">15</a>], which is also a power law model based on fetal mass, was calibrated using a different data set [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.ref025" target="_blank">25</a>]. The model of Zhang et al. [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.ref032" target="_blank">32</a>] was also calibrated using a distinct compiled data set. The Zhang et al. [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.ref032" target="_blank">32</a>] model describes kidney volume, so the units were converted assuming a tissue density of 1 g/mL. The apparently poor fit of the model of Abduljalil et al. [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0215906#pone.0215906.ref033" target="_blank">33</a>] to their own curated data set is probably due to precision-related errors—they only reported the leading coefficient of their polynomial model to one significant figure.</p>