High-resolution tandem mass spectrometry (HR-MS/MS) is a vital tool in compound identification in environmental samples, <i>e.g.</i>,  detecting unknown transformation products (TPs) which are produced when emerging contaminants are subjected to natural or anthropogenic processes. In this work the hypothesis was tested that structurally similar compounds (such as parent compounds and their TPs) have similar fragmentation patterns. Using a large set of approximately 400 compounds, the goal was to find the data evaluation approach where the similarities of known pairs was maximized. Using purchased reference standards, each compound was measured with liquid chromatography coupled to HR-MS/MS over a range of high collision dissociation (HCD) energies. Spectra were cleaned and recalibrated and are available on Norman MassBank (www.massbank.eu). TPs were paired with their respective parent compounds and included different modification reactions, such as N- or O-dealkylation, hydroxylation, or conjugation. The spectral similarity of a pair was calculated as the dot product of aligned intensity vectors. In order to find the highest similarity score for each pair, various types of HR-MS/MS spectra were compared including low and high HCDs, merged spectra, and adjusted spectra which incorporated the mass difference of the transformation. In each data evaluation approach an increase in similarity was observed when adjusted spectra were used, primarily because this led to a match of the high intensity molecular ions peaks. But the adjusted spectra in general caused lower mass fragments to no longer appear similar. The approach which yielded the highest similarity most often used merged spectra generated from the different HCD energies and included both adjusted and unadjusted fragment masses, thereby using all available information. By finding the optimum conditions for comparison this information can further be used to prioritize unknown pairs to detect new TPs in wastewater or surface water.

Towards a better understanding of spectral similarity between structurally related compounds

251st National Meeting (2016)

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