Informatics for Toxicokinetics with CvTdb and invivoPKfit
Citation:
PADILLA MERCADO, G., J. Morales, AND V. Boddu. Informatics for Toxicokinetics with CvTdb and invivoPKfit . Presented at SOT, Salt Lake City, UT, March 10 - 14, 2024. https://doi.org/10.23645/epacomptox.25400020
Impact/Purpose:
N/A
Description:
Chemical concentration versus time (CvT) data is commonly used to characterize the toxicokinetic (TK) properties of chemical substances. However, obtaining in vivo CvT data requires scientific labor and use of animal subjects. To reduce or eliminate new animal testing, it remains important to maximize reuse of existing CvT data to formulate and evaluate computational TK models. As part of this effort, the Center for Computational Toxicology and Exposure at the U.S. Environmental Protection Agency (EPA) developed a public repository of toxicokinetic data named the Concentration vs Time Database (CvTdb). CvTdb currently contains manually curated CvT data for hundreds of chemicals. CvTdb continously updates and expands as more data is extracted by our team of curators and collaborators, from the literature and publicly available resources. Data curation is an important step for efficient and effective testing of computational TK models and to identify trends in CvT data. One example of this is assessing replicability: approximately 89% of replicate observations in CvTdb data from experiments with oral or intravenous routes of administration are within two-fold of the mean concentration. Using CvTdb, we take an informatics approach to characterizing chemical TK. We developed invivoPKfit, an R package to automate the estimation of TK parameters (for 1- and 2-compartment models) from CvT data for hundreds of chemicals, integrating data across all studies available for each chemical, and quantifying uncertainty and variability in TK parameter estimates. invivoPKfit is intended to be sufficiently flexible to allow researchers to derive TK statistics from new chemical data and customized TK models. The development of both CvTdb and invivoPKfit 1) facilitates the reuse of existing CvT data, 2) provides a transparent and auditable workflow for standardized estimation of TK statistics, and 3) informs quantitative evaluation of new approach methodologies, such as TK models used for in vivo-in vitro extrapolation.
