Non-target analysis (NTA) is increasing in popularity as a comprehensive tool for the characterization of unknown chemicals,. Due to its relative novelty, new methods are regularly developed. Given the complex multi-step approaches required for NTA, even minor changes to the workflow can affect the reliability and reproducibility of the results. This is not only limited to obvious changes in methodology, such as the type of sampling, chromatographic or spectrometric settings, but also how the data is processed. Commonly, data processing approaches are compared by processing the same sample using different workflows, looking for overlap and differences between the results. This usually ignores the different sub steps that are combined within the NTA data processing, like peak picking/feature detection, componentization, filtering, prioritization, and identification. Most software has different algorithms for each of these steps, making it difficult to know where differences in the results originate from. Therefore, we investigate using a real-world dataset how the peak picking step of a workflow can potentially influence non-target results and their interpretation. Samples were collected by passive sampling in summer and winter from 2014 to 2019 at several sites across South East Queensland, Australia, providing an extensive dataset for the investigation of temporal and spatial trends. Samples were analysed using liquid chromatography coupled to an X500R high resolution mass spectrometer. Peak detection was performed using 5 different algorithms, two vendor specific (SciexOS and MarkerView) and 3 open source (MZmine 3, In House and MSDial), with parameters set as comparable as possible. The resulting feature lists were filtered, before applying further statistical analysis including principal component analysis and t-tests to find differences between the sample sets. Results vary significantly between approaches, and we examine and discuss how this could change the overall conclusion regarding spatial and temporal trends.

Effects of different detection algorithms on non-target analysis results

Where thousands of chemistry professionals meet to share ideas and advance scientific and technical knowledge. Spring 2023 abstracts, posters, and presentations will launch at the end of March.

ACS Spring 2023

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