High resolution mass spectrometry (HRMS) data has revolutionized the identification of environmental contaminants through non-targeted analyses (NTA). However, data processing and chemical identification and prioritization remain challenging due to the vast number of unknowns observed in NTA. The ideal NTA workflow requires harmonized data and tools from a variety of sources to allow the most probable and confirmed identifications. One such tool is the use of chromatographic retention time (RT). Comparing predicted RT of candidate structures to observed RT allows for additional specificity towards ultimate identification. In this work, three RT prediction models were evaluated on the same set of chemicals: 1) a logP-based model, 2) a model generated in ACD/ChromGenius, and 3) a Quantitative Structure Retention Relationship model, OPERA-RT. Our results indicate that both ACD/ChromGenius and OPERA-RT outperform the logP-based model. Between the two, OPERA-RT produced a slightly better fit on the entire set of structures than ACD/ChromGenius (R2 values of 0.85 to 0.83). Further, OPERA-RT, generated within the US EPA’s National Center for Computational Toxicology, predicted 96% of RTs within a 15% (+/-) chromatographic time window of experimental RTs. Finally, to test an NTA workflow, candidate structures were generated for formulae in the test set using the US EPA’s CompTox Chemistry Dashboard and RTs for all candidates were predicted using both ACD/ChromGenius and OPERA-RT. RT screening windows were applied to screen out unlikely candidate chemicals and enhance potential identification. Compared to ACD/ChromGenius, OPERA-RT screened out a greater percentage of the candidate structures by RT, but retained fewer of the known chemicals. This research demonstrates the potential value of including RT prediction in NTA workflows and indicates the potential value of OPERA-RT predictions to support our NTA investigations.

Evaluation of three retention time prediction models: 1) logP, 2) ACD/ChromGenius, and 3) a quantitative structure retention relationship model

254th National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.