Humans are exposed to numerous chemicals daily, for example through nutrition, therapies, and lifestyle choices, which may exert beneficial or toxicological responses. In cohort studies, exposures are frequently assessed using questionnaires, although mass spectrometry-based metabolomics has recently emerged as complementary technique capable of yielding molecular evidence of exposures. Corresponding data processing workflows, however, have been mostly developed for detecting (omnipresent) endogenous metabolites, whereas detection of exogenous chemicals would benefit from fit-for-purpose strategies. In this work, we describe novel strategies for improved exposure detection and their application to data from an untargeted metabolomics study on urine samples from the TransplantLines Food and Nutrition Biobank and Cohort Study (NCT identifier 'NCT02811835'), which includes kidney transplant recipients, potential living kidney donors, and living kidney donors (post-donation). Specifically, we describe a reference spectra generation workflow using exposure-positive samples to detect more and also previously-undetected chronic exposures, and we present a novel approach to establish detection limits based on targeted signal extraction for more reliable and lower-level detection of intermittent exposures. These approaches can contribute to unlocking additional exposure-related information from small-molecule profiling datasets thus increasing data usefulness in metabolomics research and in environmental, food, clinical, and forensic toxicology.

Original languageEnglish
Article number113188
Number of pages7
JournalFood and chemical toxicology
Publication statusPublished - Jul-2022

Cite this