You are here:
Structure identification for Non-Targeted Analytical Chemistry using the US EPA’s CompTox Chemistry Dashboard (ACS 2017 Fall meeting 3 of 3)
Citation:
McEachran, A., J. Grossman, S. Newton, K. Isaacs, K. Phillips, N. Baker, J. Sobus, Chris Grulke, AND A. Williams. Structure identification for Non-Targeted Analytical Chemistry using the US EPA’s CompTox Chemistry Dashboard (ACS 2017 Fall meeting 3 of 3). 254th American Chemical Society National Meeting, Washington, DC, August 20 - 24, 2017.
Impact/Purpose:
Presentation at the 2017 ACS Fall meeting. Identification of unknowns in non-targeted analyses (NTA) requires the integration of complementary data types to generate a confident consensus structure. These data and visualization tools indicate the capability of NTA identification within the Dashboard and demonstrate an open, accessible tool for all users of high resolution mass spectrometry (HRMS) data.
Description:
Identification of unknowns in non-targeted analyses (NTA) requires the integration of complementary data types to generate a confident consensus structure. Researchers use a variety of data and tools (e.g., chemical reference databases, spectral matching, fragment prediction tools, retention time prediction tools) to generate tentative, probable, and confirmed identifications. With the diverse, robust data contained within the US EPA’s CompTox Chemistry Dashboard, the goal of the present research is to identify and implement a harmonized identification tool and workflow using previously generated chemistry data. Data were compiled from a variety of sources, including: functional use prediction models, environmental media occurrence prediction models, and PubMed references. Scoring-based identification schemes were evaluated using compiled data on two sets of chemicals: a known test set of chemicals and a blinded, unknown mixture. Both weighted and strict scoring terms were compared to optimize identification of unknowns such that the most likely candidate structures contain the greatest scores. This approach indicates a concerted, evidence-based approach for tentative identification of unknowns without considering spectral matching and fragmentation prediction. We will also discuss development of a visualization tool in the Chemistry Dashboard, modelled after the ToxPiTM approach, wherein the relative contribution of exposure, bioactivity, and toxicity to a chemicals overall toxicity potential is depicted as slices of a pie. Applying this approach to structure identification, users can visualize the relative contribution of identification-based metrics on a list of candidate structures and observe the greatest likelihood of occurrence. These data and visualization tools indicate the capability of NTA identification within the Dashboard and demonstrate an open, accessible tool for all users of high resolution mass spectrometry (HRMS) data. This abstract does not necessarily represent the views or policies of the U.S. Environmental Protection Agency.