Lecture
Non-targeted analysis for food authentication: Tools for quality assurance and data exploration
- at -
- ICM Saal 4b
- Type: Lecture
Lecture description
For more than a decade, non-targeted analytical approaches for food authentication have been investigated in research studies. They are seen as important complement to targeted analytical methods for the verification of geographical or biological origin as wells the detection of unknown adulteration. Despite their great potential, these methods are not yet in place as accepted standard methods for official food control. Chemometrics is an important part of the non-targeted analytical workflow and its integration into the analytical perspective is relevant for bridging the gap between research and routine.
In practice, the spectra acquisition and also the classification are often wellestablished parts of the non-targeted analytical workflow, but data curation still remains a bottle neck. Therefore, we focused on the development of tools for handling non-targeted analytical data with regard to data sharing, quality assurance and data exploration. Here the tools, MONARQ (Multivariate Outlier detection of NMR data Aimed at Reporting Quality) [1] for outlier detection in non-targeted 1HNMR spectroscopy-based analysis and SERCL (flexible feature SElection, dimension Reduction & Clustering) [2] for an assumption-free way of exploring nontargeted data using graph theory algorithms will be presented.
Reference:
[1] Wustrack, F., Valentin, L., Fauhl-Hassek, C., Brockmeyer, J., Riedl, J. MONARQ - An interactive and open-source KNIME application for outlier detection in nontargeted 1H-NMR spectroscopy-based analysis, 2026, to be submitted.
[2] Nozella A., Savvateeva D., Riedl J., Vergara F., Davis A.J. SERCL - a flexible and robust pipeline for NMR chemometric feature selection, dimension reduction and clustering, 2026, to be submitted.
In practice, the spectra acquisition and also the classification are often wellestablished parts of the non-targeted analytical workflow, but data curation still remains a bottle neck. Therefore, we focused on the development of tools for handling non-targeted analytical data with regard to data sharing, quality assurance and data exploration. Here the tools, MONARQ (Multivariate Outlier detection of NMR data Aimed at Reporting Quality) [1] for outlier detection in non-targeted 1HNMR spectroscopy-based analysis and SERCL (flexible feature SElection, dimension Reduction & Clustering) [2] for an assumption-free way of exploring nontargeted data using graph theory algorithms will be presented.
Reference:
[1] Wustrack, F., Valentin, L., Fauhl-Hassek, C., Brockmeyer, J., Riedl, J. MONARQ - An interactive and open-source KNIME application for outlier detection in nontargeted 1H-NMR spectroscopy-based analysis, 2026, to be submitted.
[2] Nozella A., Savvateeva D., Riedl J., Vergara F., Davis A.J. SERCL - a flexible and robust pipeline for NMR chemometric feature selection, dimension reduction and clustering, 2026, to be submitted.
