Analytica 2026 Preview: Chemometrics for Food Quality and Authenticity

Marina Cocchi from the Department of Chemical and Geological Sciences, University of Modena and Reggio Emilia, Italy, will discuss how chemometrics can support food quality and authenticity assessment in the opening lecture of this session and will discuss Chemometrics for Food Quality and Authenticity: Approaches and Applications, starting at 15:00.

Current food concerns include nutrition, safety, sustainability, and marketing, shaped by globalized markets and regulations such as those in the European Union. Food quality is multidimensional, encompassing measurable traits as well as transparency and consumer trust. It may be influenced intentionally (for example, processing) or unintentionally (for example, contamination), while perceptions of origin and authenticity also matter. Traditional single-component analyses are often insufficient; instead, comprehensive and non-invasive approaches are needed. Chemometrics enables the extraction of meaningful information from complex data sets through exploratory analysis, visualization, and validation, providing an effective framework for food quality assessment.

Simone Lionetti from Hochschule Luzern, Switzerland, will give a lecture entitled Get Data to Clean Itself: Data Quality Audits with SelfClean, which will take place from 15:30. Accurate data is essential for reliable experimental results, yet errors such as noise, inconsistencies, and artifacts can compromise analyses. As data sets grow, manual checks and rigid rules become inadequate, requiring more systematic approaches. SelfClean addresses this by framing data cleaning as a ranking problem, enabling prioritized review or automated correction. Using self-supervised learning and distance-based metrics, it detects anomalies such as duplicates or labeling errors. Originally developed for images, the method extends to audio and potentially text data. Evaluations show strong performance with reduced manual effort. In laboratory sciences, SelfClean could support validation of sensor-based data like spectra or chromatograms, combining AI with human expertise to improve data quality.

Janet Riedl from the German Federal Institute for Risk Assessment (BfR), Germany, will present on Non-Targeted Analysis for Food Authentication: Tools for Quality Assurance and Data Exploration starting at 16:00. Non-targeted methods have long complemented targeted analyses for verifying food origin and detecting adulteration, but they are not yet routine in official control. Integrating chemometrics is key to making them practical. While data acquisition is well established, managing and curating data sets remains challenging. This lecture introduces tools to address these issues: MONARQ for outlier detection in ¹H-NMR data and SERCL for flexible, assumption-free exploration using graph-based methods. These tools support data sharing, quality assurance, and deeper exploration, helping translate non-targeted approaches into routine food authentication.

Jochen Vestner from the Institute for Viticulture and Oenology at the Dienstleistungszentrum Ländlicher Raum Rheinpfalz – Weincampus, Germany, will present on Automated Non-Targeted GC–MS Data Analysis, scheduled from 16:30. Non-targeted analysis is widely used in metabolomics, environmental, and food research, providing an unbiased view of known and unknown compounds. Although LC–HRMS is common, GC–MS remains valuable due to strong separation performance and accessible instrumentation. However, data processing steps like baseline correction and alignment are often complex and manual. This lecture presents automated supervised and unsupervised approaches that avoid traditional feature detection. By segmenting chromatograms, applying transformations, and using multiway decomposition, these methods streamline analysis. A browser-based, open-source Python application makes the supervised workflow accessible to a wider audience.