You are here:
A Method to Quantify Reproducibility in PBPK Model Methods and Results
Citation:
Sayre, R., Woodrow Setzer, J. Wambaugh, AND Chris Grulke. A Method to Quantify Reproducibility in PBPK Model Methods and Results . Presented at OpenTox, RTP, NC, July 11 - 12, 2018. https://doi.org/10.23645/epacomptox.7080740
Impact/Purpose:
Poster presentation at OpenTox meeting July 2018
Description:
Physiologically-based pharmacokinetic (PBPK) models estimate chemical concentration changes in different body organs using a system of equations representing relevant biological processes and physicochemical properties. PBPK models are key to relating biochemical responses measured in assays to administered doses in whole-animal toxicity studies for high-throughput hazard assessment. While building a public database of 973 published PBPK models as a resource for risk assessors and modelers, we developed a minimum standard for reproducibility represented by eleven binary classifications of documentation, mathematical fidelity, and predictive power. Few models supported methods reproducibility: 49.7% met our documentation criteria (with the most common failure being incomplete description of parameter values), and only 20.7% of the papers (10.3% of the total set) met the mathematical basis review criteria (frequently because of inconsistent units in the equations). In our results reproducibility assessment, 28.2% of the total set of papers did not present adequate experimental data to justify the validity of the model and only 2.6% of the papers provided model code. We re-ran the models for all papers with code, and none of the published code perfectly reproduced the corresponding figures from its paper. Mathematical errors in code, as well as inconsistencies between paper and code parameter values, were noted. Our technical review process efficiently evaluates PBPK methods and results reproducibility to validate consideration of a model’s use for chemical risk decision making. We propose our manuscript evaluation method as a first step for PBPK model review until a curated model database becomes available. This abstract does not necessarily reflect U.S. EPA policy.
URLs/Downloads:
DOI: A Method to Quantify Reproducibility in PBPK Model Methods and Results
RRSAYRENCOPENTOXPOSTER01_FINAL.PDF (PDF, NA pp, 389.543 KB, about PDF)