Case Study 7


Facilitated by Prachi Pradeep and Richard Judson (National Center for Computational Toxicology, ORD, US EPA) This does not reflect EPA policy.

Goal: Risk assessment and categorization of 21 Hindered Phenols (HP).  One of the issues is to investigate whether particular HPs have the potential to be estrogenic or not, and if so, their relative potency using read-across and/or (Q)SAR methods.

The aim is to utilize physchem properties, data quality measures and available HTS data streams in conjunction with various structure descriptor methods to identify analogs, to develop categories and to evaluate the advantages and uncertainties associated with read-across techniques.

Challenge: Potential source analogs for read-across are typically identified based on structural similarity. Although much guidance has been published for read-across, practical principles for the identification and evaluation of the scientific validity of source analogs remains lacking. Additionally, it is not clear what is the best descriptor set. Finally, the data used to train and evaluate any models is noisy.

Data set: The available data set comprises 462 hindered phenols and 257 non-hindered phenols with ER binding data. Here we have a rough metric of data quality, allowing modelers to exclude data with varying levels of quality.