You are here:
Quantitative structure-property relationships for chemical functional use and weight fractions in consumer articles
Phillips, K., J. Wambaugh, P. Egeghy, C. Grulke, AND K. Isaacs. Quantitative structure-property relationships for chemical functional use and weight fractions in consumer articles. Society of Toxicology Annual Meeting, New Orleans, LA, March 13 - 17, 2016.
This research details a new approach for predicting the functional role of data-poor chemicals from structure and property information for use in exposure and risk prediction.
Chemical functional use -- the functional role a chemical plays in processes or products -- may be a useful heuristic for predicting human exposure potential in that it comprises information about the compound's likely physical properties and the product formulations or articles in which it is likely to be found. Functional use is also an integral part of alternatives assessment, which involves identifying safer chemicals that can perform the same role in products as current chemicals. Using publicly available information, we have compiled a dataset of 14,000 chemicals and 240 functional uses. This dataset was used to build classifier models for 49 functional use categories using chemical structures and physical properties which were described using ToxPrint chemical descriptors and the US Environmental Protection Agency's Estimation Program Interface Suite, respectively. The models developed here will be useful in developing and parameterizing high-throughput models of human exposures for chemicals for which little measured or reported data are available.