Rapid detection of chemicals of emerging concern from non-target High-Resolution Mass Spectrometry

Nontarget High-Resolution Mass Spectrometry (HRMS) is becoming a key method for detecting chemical exposure in human and environmental systems. However, variations in vendor-specific data files and specialized software create challenges in comparing samples and detecting unknown compounds. Current detection methods often rely on mass defect, homologous patterns, and sample-dependent approaches, which are labour-intensive and limited.

A vendor-independent, software was developed to automates nontarget HRMS processing, enabling the identification known compounds and the generation of suspect lists for unidentified compounds. Additionally, a machine learning algorithm was created to detect unidentified Per- and polyfluoroalkyl substances(PFAS) from non-target data, along side an algorithm to match unidentified compounds using their spectral fingerprints. These algorithms were applied to the NORMANews dataset, a global collection of hundreds of samples from various matrices to identify potential PFAS.

Please note this is a Student PhD Progress review presentation by Mathieu Feraud.