Improving the annotation of post-prandial metabolites in human urine samples from an untargeted LC-MS-based bilberry-blueberry intervention study by Feature-Based Molecular Networking

This research aims to extend the annotation coverage of metabolites occurring in human urine samples from an already published intervention study on bilberry (V. myrtillus) and blueberry (V. corymbosum) intake, combining LC-MS standard feature detection tools with MN, i.e. Feature-Based Molecular Networking (FBMN). By this approach, using the GNPS infrastructure including public spectral libraries, 66 additional potential marker of berries intake were annotated, increasing the annotation coverage of the 45% compared to the previous research, and the structures of fifteen additional metabolites were hypothesized by spectral analysis. Moreover, several molecular families of the phase II (e.g. glucuronidated and sulfated phenolics) and phase I (e.g. phenylpropionic acids) metabolism were identified by longitudinal correlation analysis. We expect that our approach will greatly assist in disentangling the microbial and human biotransformations of food-derived metabolites.