You are here:
Using Prepared Mixtures of ToxCast Chemicals to Evaluate Non-Targeted Analysis (NTA) Method Performance.
Citation:
Sobus, J., J. Grossman, A. Chao, R. Singh, A. Williams, Chris Grulke, A. Richard, S. Newton, A. McEachran, AND E. Ulrich. Using Prepared Mixtures of ToxCast Chemicals to Evaluate Non-Targeted Analysis (NTA) Method Performance. Analytical and Bioanalytical Chemistry. Springer, New York, NY, 411(4):835-851, (2019). https://doi.org/10.1007/s00216-018-1526-4
Impact/Purpose:
This research article highlights findings from a non-targeted analysis (NTA) study of 10 synthetic mixtures prepared using 1,269 ToxCast chemicals. Results of the analysis demonstrate the suitability of NTA methods for detecting ToxCast chemicals in complex mixtures. Our findings should be considered by any researcher, manager, or regulator looking to make science or policy decisions based on NTA data.
Description:
Non-targeted analysis (NTA) methods are increasingly used to discover chemicals of emerging concern (CECs). The extent to which these methods can support exposure and health studies must now be determined. EPA’s Non-Targeted Analysis Collaborative Trial (ENTACT) was launched in 2016 to address this need. As part of ENTACT, 1,269 unique substances from EPA’s ToxCast library were combined to make ten synthetic mixtures, with each mixture containing between 95 and 365 substances. Blinded NTA was first performed on each mixture using liquid chromatography (LC) coupled with high-resolution mass spectrometry (HRMS). An unblinded evaluation was then performed to identify limitations of the NTA method. Overall, 60% of spiked substances could be observed and 40% were not observed. Discounting spiked isomers, true positive rates from the blinded analysis ranged from 27% to 51%. True positive rates from the unblinded evaluation reached a maximum of 68%. Reproducibility rates averaged 75% for substances that were spiked into more than one mixture and observed at least once. Considerable discordance in substance identification was observed when comparing a subset of results derived from two separate chromatography methods. We conclude that a single NTA method, even when optimized, can likely characterize only a subset of ToxCast substances (and, by extension, other CECs). Rigorous quality control and self-evaluation practices should be a requirement of labs generating NTA data to support exposure and health studies. Accurate and transparent communication of performance results will best enable meaningful interpretations of NTA data, and defensible decisions by those wishing to use these data.
URLs/Downloads:
DOI: Using Prepared Mixtures of ToxCast Chemicals to Evaluate Non-Targeted Analysis (NTA) Method Performance.