Strategy for comparative untargeted metabolomics reveals honey markers of different floral and geographic origins using ultrahigh-performance liquid chromatography-hybrid quadrupole-orbitrap mass spectrometry

- A comparative untargeted metabolomics strategy using UHPLC-Q-Orbitrap was proposed.
- The data were processed using a three-stage approach to ensure marker authenticity.
- Marker structure was elucidated by correlation between molecules and fragment ions.
- A mass accuracy was less than 1.0 ppm for protonated molecules and fragment ions.
- Marker of honey samples from various floral and geographic origins was identified.

Honey discrimination based on floral and geographic origins is limited by the ability to determine reliable markers because developing hypothetical substances in advance considerably limits the throughput of metabolomics studies. Here, we present a novel approach to screen and elucidate honey markers based on comparative untargeted metabolomics using ultrahigh-performance liquid chromatography-hybrid quadrupole-orbitrap mass spectrometry (UHPLC-Q-Orbitrap). To reduce metabolite information losses during sample preparation, the honey samples were dissolved in water and centrifuged to remove insoluble particles prior to UHPLC-Q-Orbitrap analysis in positive and negative electrospray ionization modes. The data were pretreated using background subtraction, chromatographic peak extraction, normalization, transformation and scaling to remove interferences from unwanted biases and variance in the experimental data. The pretreated data were further processed using principal component analysis (PCA) and a three-stage approach (t-test, volcano plot and variable importance in projection (VIP) plot) to ensure marker authenticity. A correlation between the molecular and fragment ions with a mass accuracy of less than 1.0 ppm was used to annotate and elucidate the marker structures, and the marker responses in real samples were used to confirm the effectiveness of the honey discrimination. Moreover, we evaluated the data quality using blank and quality control (QC) samples based on PCA clustering, retention times, normalized levels and peak areas. This strategy will help guide standardized, comparative untargeted metabolomics studies of honey and other agro-products from different floral and geographic origins.