You are here:
ISES Talk: Rapid Parameterization of Pathway Specific Exposure Models
Citation:
Isaacs, K., K. Phillips, J. Wambaugh, P. Price, AND K. Dionisio. ISES Talk: Rapid Parameterization of Pathway Specific Exposure Models. ISES-ISEE 2018, Ottawa, CANADA, August 26 - 30, 2018.
Impact/Purpose:
U.S. EPA, under its ExpoCast program, is developing high-throughput near-field modeling methods to estimate chemical exposure and to provide real-world context to high-throughput screening (HTS) data. This abstract describes progress made in parameterizing multiple pathway-specific exposure model, thereby enabling consensus modeling across multiple pathways.
Description:
Evaluation of biomonitoring data has shown the importance of near-field chemical exposure pathways. New high throughput (HT) exposure models for consumer products and foods have been developed, but require large amounts of data on the composition of consumer products, product use, and exposure factor data. Under EPA’s ExpoCast program, we have led a large effort to obtain and organize publicly-available information for thousands of chemicals to support these models. However, challenges can arise in efficiently evaluating and employing large amounts of data for a novel application (i.e., a new model, chemical, product, use, or population). Thoughtful model design and data organization promotes rapid assessments of new products, as does the establishment of formal linkages among chemicals, inputs, and algorithms describing exposure processes. For example, efficient modeling of chemicals in consumer products is enabled by a fit-for-purpose system of consumer product categories. Categories can be linked to generic exposure scenarios that define indoor fate and transport, route-specific intake, or disposal processes and associated product specific factors (e.g., use and release patterns) thereby creating a library of fully parameterized algorithms. In this system, obtaining results for a new chemical does not require a de novo parameterization, only knowledge of its concentrations in products. We plan to build upon this generalization to obtain exposure results for generic chemicals within categories (e.g., via Latin hypercube sampling of the physicochemical space of chemicals in products) allowing for rapid read-across of exposure results. Finally, while near-field models may adequately capture chemical-to-chemical variability in population median exposure, new HT models of ambient exposures (e.g., associated with industrial releases) or occupational exposures will be required to address populations having outlier exposure patterns. Obtaining and organizing available data and algorithms for these pathways is a current area of focus in ExpoCast.
