You are here:
Rapid collection of experimental physicochemical property data to inform various models and testing methods
Nicolas, C., K. Mansouri, K. Phillips, Chris Grulke, A. Richard, A. Williams, J. Rabinowitz, K. Isaacs, A. Yau, AND J. Wambaugh. Rapid collection of experimental physicochemical property data to inform various models and testing methods. American Chemical Society Annual Meeting, New Orleans,LA, March 18 - 22, 2018.
This was a presentation to the 255th American Chemical Society annual meeting in New Orleans, LA.
In order to determine the potential toxicological effects, toxicokinetics, and route(s) of exposure for chemicals, their structures and corresponding physicochemical properties are required. With this data, the risk for thousands of environmental chemicals can be prioritized. However, as there are limitations on the availability on experimental data, we have attempted to efficiently fill these data gaps by generating new data for 200 structurally diverse compounds. This set of compounds were rigorously selected from the USEPA Distributed Structure-Searchable Toxicity Database (DSSTox). Evaluated in this pilot study are rapid experimental methods to determine five physicochemical properties including the log of the octanol:water partition coefficient (log(Kow)), vapor pressure, water solubility, Henry’s law constant, and the acid dissociation constant. For the majority of the compounds in this study, experiments were successful for at least one property, with log(Kow) yielding the largest return (176 values). Using ToxPrint Chemotypes, it was observed that the presence of 21 structural features may have played an overall role in rapid estimation failures. Where available, the new estimates were compared with previous measurements in order to illustrate the consistency of new experimental data with traditional measurement methods. As quantitative structure-property relationship (QSAR) models are relied upon for filling huge gaps in physicochemical property information, we evaluated 5 suites of QSARs for their predictive ability and chemical coverage or applicability domain of the new experimental estimates. Accurate measurements of these properties are crucial for facilitating better exposure predictions in two ways: 1) direct parameterization of exposure models; and 2) construction of physicochemical property QSARs with a wider applicability domain, whereby their resulting predictions can be used to parameterize exposure models in the absence of experimental data.