You are here:
Using Web-based Interspecies Correlation Estimation (Web-ICE) models as a tool for acute toxicity prediction
Citation:
Willming, M., C. Lilavois, Sandy Raimondo, AND M. Barron. Using Web-based Interspecies Correlation Estimation (Web-ICE) models as a tool for acute toxicity prediction. SETAC North America 36th Annual Meeting, Salt Lake City, UT, November 01 - 05, 2015.
Impact/Purpose:
Abstract summarizing ICE models and Web-ICE
Description:
In order to assess risk of contaminants to taxa with limited or no toxicity data available, Interspecies Correlation Estimation (ICE) models have been developed by the U.S. Environmental Protection Agency to extrapolate contaminant sensitivity predictions based on data from commonly tested species. ICE models are log-linear least squares regressions that estimate acute toxicity (LC50/LD50) to an untested species, genus, or family (predicted taxon) from known toxicity of a single surrogate species. Approximately 2400 ICE models have currently been developed and validated. Databases are available for aquatic animals, algae, and wildlife using the online Web-ICE tool (http://epa.gov/ceampubl/fchain/webice/), which also includes modules for endangered species extrapolation and generation of species sensitivity distributions (SSDs). Web-ICE toxicity estimates and hazard concentration predictions have demonstrated high accuracy for closely related taxa, and models can be used for threatened and endangered species where toxicity data are limited. This presentation will provide an overview of the Web-ICE tool and approaches for selection of robust models as well as updates on current work improving model uncertainly and prediction accuracy.