Stephan Freitag
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Stephan Freitag has a total of 84 co-authors. Here are the top co-authors by number of shared publications:
- Rudolf Krska
Department of Agricultural Sciences, Institute of Bioanalytics and Agro-Metabolomics, BOKU University, Tulln, Austria
17 shared publications - Michael Sulyok
Department of Agricultural Sciences, Institute of Bioanalytics and Agro-Metabolomics, BOKU University, 3430 Tulln, Austria
9 shared publications - Boris Mizaikoff
Institute of Analytical and Bioanalytical Chemistry Ulm University Ulm Germany
6 shared publications - Dedy Leonardo Nadeak
BOKU University
6 shared publications - Polina Fomina
Institute of Analytical and Bioanalytical Chemistry Ulm University Ulm Germany
5 shared publications Achim Kohler
Faculty of Science and Technology, Norwegian University of Life Sciences, 1432 Ås, Norway
4 shared publications
Latest Publications
- Open Access
- 12 August 2026
- Published by Springer Nature in Analytical and Bioanalytical Chemistry
- Vol. 418 (23) , 7903-7912
Discrimination of plant-based milk alternatives: a pilot study comparing mid and near-infrared spectroscopy
- 12 August 2026
- Published by Springer Nature in Analytical and Bioanalytical Chemistry
- Vol. 418 (23) , 7903-7912
Open AccessDiscrimination of plant-based milk alternatives: a pilot study comparing mid and near-infrared spectroscopy
- Open Access
- 19 July 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Model Population Analysis to Optimize Near-Infrared-Based Screening of Deoxynivalenol in Wheat
The widespread contamination of wheat with deoxynivalenol (DON) necessitates rapid, non-destructive screening methods. Although near-infrared (NIR) spectroscopy offers a promising cost-effective solution, indirect detection via Fusarium-induced sample compositional changes often suffers from high model variance and overfitting. This study systematically evaluated the impact of seven methodological factors, namely dataset selection, dataset balancing, sample grinding, toxin type, sample weighting, spectral preprocessing, and variable selection, using model population analysis (MPA) to assess the performance of partial least squares discriminant analysis (PLS-DA) sub-models for classifying wheat samples as compliant or non-compliant with the European Commission (EC) limit of 1000 μg/kg DON. By using repeated random splits with a stratification constraint across the seven methodological factors, 12,800 PLS-DA sub-models were generated. To study the effects of each factor, a generalized linear mixed model (GLMM) was employed. The results indicated that model performance was highly influenced by all methodological factors, with the toxin type and dataset balancing showing the greatest impacts on the F1-score and True Positive Rate (TPR), respectively. The optimal pipeline achieved an estimated F1-score of 0.74 ± 0.01 and a TPR of 0.84 ± 0.01 by utilizing Austrian-only wheat samples, intact kernels, standard normal variate (SNV) preprocessing, and downsampling for data balancing, without variable selection. Notably, samples labelling based on total DON (tDON, sum of DON and its acetylated and modified forms) improved classification performance, indicating that there might be a stronger correlation with Fusarium-induced sample compositional changes with tDON than with DON. Furthermore, applying an observation-weighting approach to highly uncertain samples mitigated the propagation of reference analytical error, significantly reducing false negative misclassifications near the regulatory threshold. Overall, our results provide a rigorously validated framework for optimizing complex chemometric pipelines, establishing a robust NIR methodology for challenging application such as indirect non-destructive screening of DON in wheat.
- 19 July 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessModel Population Analysis to Optimize Near-Infrared-Based Screening of Deoxynivalenol in Wheat
The widespread contamination of wheat with deoxynivalenol (DON) necessitates rapid, non-destructive screening methods. Although near-infrared (NIR) spectroscopy offers a promising cost-effective solution, indirect detection via Fusarium-induced sample compositional changes often suffers from high model variance and overfitting. This study systematically evaluated the impact of seven methodological factors, namely dataset selection, dataset balancing, sample grinding, toxin type, sample weighting, spectral preprocessing, and variable selection, using model population analysis (MPA) to assess the performance of partial least squares discriminant analysis (PLS-DA) sub-models for classifying wheat samples as compliant or non-compliant with the European Commission (EC) limit of 1000 μg/kg DON. By using repeated random splits with a stratification constraint across the seven methodological factors, 12,800 PLS-DA sub-models were generated. To study the effects of each factor, a generalized linear mixed model (GLMM) was employed. The results indicated that model performance was highly influenced by all methodological factors, with the toxin type and dataset balancing showing the greatest impacts on the F1-score and True Positive Rate (TPR), respectively. The optimal pipeline achieved an estimated F1-score of 0.74 ± 0.01 and a TPR of 0.84 ± 0.01 by utilizing Austrian-only wheat samples, intact kernels, standard normal variate (SNV) preprocessing, and downsampling for data balancing, without variable selection. Notably, samples labelling based on total DON (tDON, sum of DON and its acetylated and modified forms) improved classification performance, indicating that there might be a stronger correlation with Fusarium-induced sample compositional changes with tDON than with DON. Furthermore, applying an observation-weighting approach to highly uncertain samples mitigated the propagation of reference analytical error, significantly reducing false negative misclassifications near the regulatory threshold. Overall, our results provide a rigorously validated framework for optimizing complex chemometric pipelines, establishing a robust NIR methodology for challenging application such as indirect non-destructive screening of DON in wheat.
- Open Access
- 15 June 2026
- Published by American Chemical Society (ACS) in ChemRxiv
The environmental impact of mycotoxin analysis: Evaluating the greenness of liquid chromatography mass spectrometry
Food safety relies on extensive analytical measurements, while the environmental impact caused remains unclear. In this study, we evaluated the greenness of liquid chromatography–mass spectrometry (LC–MS), the analytical workhorse for mycotoxin screening, using the AGREE metric paired with a carbon footprint (CF) estimate for instrument operation and solvent use. Using manufacturer specifications for a routine setup, we estimate that a single system emits 30 t CO2 per year at 8,000 samples, with 6 t arising from the instrument being on standby, as operated by authors in their laboratory. We show that electricity consumption dominates the CF, whereas solvent consumption contributes 2% to the estimated total CF. Supplying the same workload with solar electricity can lowers annual emissions to ~4 t (≈86% reduction), emphasizing the usage of decarbonized power. Only a moderate influence of multi analyte LC-MS based screening on the AGREE metric was found. Contrary to that the CF estimation showed that increasing the number of analytes per analysis can markedly reduce emissions. We show that LC-MS sustainability can be improved through method optimization by expanding the number of target analytes, shortening analysis times, and reducing system stand-by, when screening food and feed for residues or contaminants such as mycotoxins. In this way, analytical laboratories can already lower their environmental footprint without investing into new equipment or switching to greener solvents.
- 15 June 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessThe environmental impact of mycotoxin analysis: Evaluating the greenness of liquid chromatography mass spectrometry
Food safety relies on extensive analytical measurements, while the environmental impact caused remains unclear. In this study, we evaluated the greenness of liquid chromatography–mass spectrometry (LC–MS), the analytical workhorse for mycotoxin screening, using the AGREE metric paired with a carbon footprint (CF) estimate for instrument operation and solvent use. Using manufacturer specifications for a routine setup, we estimate that a single system emits 30 t CO2 per year at 8,000 samples, with 6 t arising from the instrument being on standby, as operated by authors in their laboratory. We show that electricity consumption dominates the CF, whereas solvent consumption contributes 2% to the estimated total CF. Supplying the same workload with solar electricity can lowers annual emissions to ~4 t (≈86% reduction), emphasizing the usage of decarbonized power. Only a moderate influence of multi analyte LC-MS based screening on the AGREE metric was found. Contrary to that the CF estimation showed that increasing the number of analytes per analysis can markedly reduce emissions. We show that LC-MS sustainability can be improved through method optimization by expanding the number of target analytes, shortening analysis times, and reducing system stand-by, when screening food and feed for residues or contaminants such as mycotoxins. In this way, analytical laboratories can already lower their environmental footprint without investing into new equipment or switching to greener solvents.
- Open Access
- 1 May 2026
- Published by Elsevier in Food Control
- Vol. 188
Mycotoxin field trial uncovers first aflatoxin B1 occurrence in maize in Lower Austria
- 1 May 2026
- Published by Elsevier in Food Control
- Vol. 188
Open AccessMycotoxin field trial uncovers first aflatoxin B1 occurrence in maize in Lower Austria
- Open Access
- 21 January 2026
- Published by American Chemical Society (ACS) in Analytical Chemistry
- Vol. 98 (4) , 2775-2782
Mid-Infrared Spectroscopy for the Qualitative and Quantitative Analysis of the Wheat Proteome
Wheat contributes 19% of protein in the global human diet. Mid-infrared spectroscopy (MIR) combined with attenuated total reflection (ATR) is an excellent analytical method for assessing the chemical composition of complex biological samples. In contrast to established techniques used for wheat protein analysis, we show that by using ATR-MIR the protein secondary structures of different fractions can be studied. We found that albumin and globulin fractions were primarily composed of α-helix, with proportions of 57.8% and 45.9%, respectively. Gliadins, meanwhile, contained 38.3% β-turn and 36.9% α-helix, while glutenins predominantly exhibited 44.8% β-turn secondary structures. In addition, we found that by using analysis of variance (ANOVA) simultaneous component analysis (ASCA), the obtained MIR spectra of the wheat protein fractions were significantly (p < 0.001) affected by the sampling sites and variety. Quantification of the protein content of each sample through the amide II band also revealed significant differences (p < 0.001) across the sampling sites in the different protein fractions. The found concentration range of protein fractions within the wheat samples were as follows: 1.7-3.6, 0.4-2.4, 4.0-5.4, and 0.8-4.8 g/100 g of albumins, globulins, gliadins, and glutenins, respectively. In conclusion, this study shows that ATR-MIR has an immense potential for wheat proteome analysis.
- 21 January 2026
- Published by American Chemical Society (ACS) in Analytical Chemistry
- Vol. 98 (4) , 2775-2782
Open AccessMid-Infrared Spectroscopy for the Qualitative and Quantitative Analysis of the Wheat Proteome
Wheat contributes 19% of protein in the global human diet. Mid-infrared spectroscopy (MIR) combined with attenuated total reflection (ATR) is an excellent analytical method for assessing the chemical composition of complex biological samples. In contrast to established techniques used for wheat protein analysis, we show that by using ATR-MIR the protein secondary structures of different fractions can be studied. We found that albumin and globulin fractions were primarily composed of α-helix, with proportions of 57.8% and 45.9%, respectively. Gliadins, meanwhile, contained 38.3% β-turn and 36.9% α-helix, while glutenins predominantly exhibited 44.8% β-turn secondary structures. In addition, we found that by using analysis of variance (ANOVA) simultaneous component analysis (ASCA), the obtained MIR spectra of the wheat protein fractions were significantly (p < 0.001) affected by the sampling sites and variety. Quantification of the protein content of each sample through the amide II band also revealed significant differences (p < 0.001) across the sampling sites in the different protein fractions. The found concentration range of protein fractions within the wheat samples were as follows: 1.7-3.6, 0.4-2.4, 4.0-5.4, and 0.8-4.8 g/100 g of albumins, globulins, gliadins, and glutenins, respectively. In conclusion, this study shows that ATR-MIR has an immense potential for wheat proteome analysis.
- Open Access
- 22 December 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Mycotoxin field trial uncovers first Aflatoxin B1 occurrence in maize in Lower Austria
Climate change is driving the migration of crop pests and pathogens into new regions, posing significant threats to food security and safety. Aflatoxins, toxic metabolites produced by Aspergillus flavus, are of particular concern due to their carcinogenic properties. This study reports the first detection of aflatoxin B1 (AFB1) in maize grown in Tulln, Lower Austria, a region where such contamination has not been previously observed. The data were collected as part of the BOKU TTox-Monitor trial, a long-term monitoring trial to study the impact of climate change on the (co-)occurrence of mycotoxins. In 2024, 24 maize varieties were cultivated and analysed for fungal metabolites using liquid chromatog-raphy-tandem mass spectrometry (LC-MS/MS). AFB1 was detected in one plot at concentrations ranging from 4.47 to 29.49 µg/kg, and the presence of A. flavus was confirmed through mycological analysis. The unusually hot and dry condi-tions in 2024, coupled with severe corn borer infestation, likely created an optimal environment for A. flavus infection and aflatoxin production. While most samples were below regulatory limits for other mycotoxins, significant variability in contamination levels was observed across varieties. This finding supports predictions that aflatoxin contamination will become a growing concern in central Europe as climate change progresses, and underscores the need for long-term monitoring, research, and mitigation strategies to address this emerging food safety threat.
- 22 December 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessMycotoxin field trial uncovers first Aflatoxin B1 occurrence in maize in Lower Austria
Climate change is driving the migration of crop pests and pathogens into new regions, posing significant threats to food security and safety. Aflatoxins, toxic metabolites produced by Aspergillus flavus, are of particular concern due to their carcinogenic properties. This study reports the first detection of aflatoxin B1 (AFB1) in maize grown in Tulln, Lower Austria, a region where such contamination has not been previously observed. The data were collected as part of the BOKU TTox-Monitor trial, a long-term monitoring trial to study the impact of climate change on the (co-)occurrence of mycotoxins. In 2024, 24 maize varieties were cultivated and analysed for fungal metabolites using liquid chromatog-raphy-tandem mass spectrometry (LC-MS/MS). AFB1 was detected in one plot at concentrations ranging from 4.47 to 29.49 µg/kg, and the presence of A. flavus was confirmed through mycological analysis. The unusually hot and dry condi-tions in 2024, coupled with severe corn borer infestation, likely created an optimal environment for A. flavus infection and aflatoxin production. While most samples were below regulatory limits for other mycotoxins, significant variability in contamination levels was observed across varieties. This finding supports predictions that aflatoxin contamination will become a growing concern in central Europe as climate change progresses, and underscores the need for long-term monitoring, research, and mitigation strategies to address this emerging food safety threat.
- Open Access
A portable 3D-printed near-infrared spectrometer to screen maize for deoxynivalenol and zearalenone contamination
Mycotoxins are toxic metabolites found in grains and cereals, posing a severe potential risk to human and animal health and significantly disrupting animal production, particularly in industries like pig farming. Current laboratory methods for mycotoxin analysis are expensive and time-intensive, creating a need for faster, more accessible solutions. This study set out to address this challenge by developing a portable spectrometer prototype based on the near-infrared (NIR) region and designing a chemometric model to classify ground maize as either non-compliant (NC) or compliant (C) for zearalenone (ZON) and deoxynivalenol (DON). A total of 259 naturally contaminated maize samples, collected over two years from diverse European regions, were analyzed using both liquid chromatography-tandem mass spectrometry (LC-MS/MS) and the prototype device. Spectral data were preprocessed using the first derivative, and a partial least squares discriminant analysis (PLS-DA) model was developed to classify ZON and DON levels into NC and C categories. Thresholds of 100 µg/kg for ZON and 500 µg/kg for DON were used to define compliance. The PLS-DA model showed good performance for ZON, achieving a classification accuracy of 86.3%. However, for DON classification a rather limited accuracy of 66.75% was achieved. While the DON model could identify NC samples it struggled with C samples. Despite these challenges, the results highlight the potential of a portable NIR spectrometer, combined with straightforward preprocessing and PLS-DA modeling, as a rapid, cost-effective screening tool for detecting mycotoxins in maize. The presented approach could significantly simplify mycotoxin monitoring, offering a practical solution to safeguard public health and enhance agricultural productivity.
Open AccessA portable 3D-printed near-infrared spectrometer to screen maize for deoxynivalenol and zearalenone contamination
Mycotoxins are toxic metabolites found in grains and cereals, posing a severe potential risk to human and animal health and significantly disrupting animal production, particularly in industries like pig farming. Current laboratory methods for mycotoxin analysis are expensive and time-intensive, creating a need for faster, more accessible solutions. This study set out to address this challenge by developing a portable spectrometer prototype based on the near-infrared (NIR) region and designing a chemometric model to classify ground maize as either non-compliant (NC) or compliant (C) for zearalenone (ZON) and deoxynivalenol (DON). A total of 259 naturally contaminated maize samples, collected over two years from diverse European regions, were analyzed using both liquid chromatography-tandem mass spectrometry (LC-MS/MS) and the prototype device. Spectral data were preprocessed using the first derivative, and a partial least squares discriminant analysis (PLS-DA) model was developed to classify ZON and DON levels into NC and C categories. Thresholds of 100 µg/kg for ZON and 500 µg/kg for DON were used to define compliance. The PLS-DA model showed good performance for ZON, achieving a classification accuracy of 86.3%. However, for DON classification a rather limited accuracy of 66.75% was achieved. While the DON model could identify NC samples it struggled with C samples. Despite these challenges, the results highlight the potential of a portable NIR spectrometer, combined with straightforward preprocessing and PLS-DA modeling, as a rapid, cost-effective screening tool for detecting mycotoxins in maize. The presented approach could significantly simplify mycotoxin monitoring, offering a practical solution to safeguard public health and enhance agricultural productivity.
- Open Access
- 14 August 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Mid-Infrared Spectroscopy for the Qualitative and Quantitative Analysis of the Wheat Proteome
Wheat contributes 19% of the protein in the global human diet. Mid-infrared spectroscopy (MIR) combined with attenuated total reflection (ATR) is an excellent analytical method for assessing the chemical composition of complex biological samples. In contrast to established techniques used for wheat protein analysis, we show that by using ATR-MIR, the protein secondary structures of different fractions can be studied. We found that albumin and globulin fractions were primarily composed of α-helix, with proportions of 57.8% and 45.9%, respectively. Gliadins, meanwhile, contained 38.3% β-turn and 36.9% α-helix, while glutenins predominantly exhibited 44.8% β-turn secondary structures. In addition, we found that by using analysis of variance (ANOVA) simultaneous component analysis (ASCA), the obtained MIR spectra of the wheat protein fractions were significantly (p < 0.001) affected by the sampling sites and variety. Quantification of the protein content of each sample through the amide II band also revealed significant differences (p < 0.001) across the sampling sites in the different protein fractions. The found concentration range of protein fractions within the wheat samples were as follows: 1.7 – 3.6, 0.4 – 2.4, 4.0 – 5.4, and 0.8 – 4.8 g/100 g of albumins, globulins, gliadins, and glutenins, respectively. In conclusion, this study shows that ATR-MIR has an immense potential for wheat proteome analysis.
- 14 August 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessMid-Infrared Spectroscopy for the Qualitative and Quantitative Analysis of the Wheat Proteome
Wheat contributes 19% of the protein in the global human diet. Mid-infrared spectroscopy (MIR) combined with attenuated total reflection (ATR) is an excellent analytical method for assessing the chemical composition of complex biological samples. In contrast to established techniques used for wheat protein analysis, we show that by using ATR-MIR, the protein secondary structures of different fractions can be studied. We found that albumin and globulin fractions were primarily composed of α-helix, with proportions of 57.8% and 45.9%, respectively. Gliadins, meanwhile, contained 38.3% β-turn and 36.9% α-helix, while glutenins predominantly exhibited 44.8% β-turn secondary structures. In addition, we found that by using analysis of variance (ANOVA) simultaneous component analysis (ASCA), the obtained MIR spectra of the wheat protein fractions were significantly (p < 0.001) affected by the sampling sites and variety. Quantification of the protein content of each sample through the amide II band also revealed significant differences (p < 0.001) across the sampling sites in the different protein fractions. The found concentration range of protein fractions within the wheat samples were as follows: 1.7 – 3.6, 0.4 – 2.4, 4.0 – 5.4, and 0.8 – 4.8 g/100 g of albumins, globulins, gliadins, and glutenins, respectively. In conclusion, this study shows that ATR-MIR has an immense potential for wheat proteome analysis.
- Open Access
- 4 May 2025
- Published by Elsevier in Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy
- Vol. 340, 126340
Attenuated total reflection mid-infrared spectroscopy to screen Austrian and French wheat from multiple years for deoxynivalenol
- 4 May 2025
- Published by Elsevier in Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy
- Vol. 340, 126340
Open AccessAttenuated total reflection mid-infrared spectroscopy to screen Austrian and French wheat from multiple years for deoxynivalenol
- Open Access
- 8 January 2025
- Published by American Chemical Society (ACS) in ACS Food Science & Technology
- Vol. 5 (1) , 209-217
Innovative Infrared Spectroscopic Technologies for the Prediction of Deoxynivalenol in Wheat
Mycotoxin contamination in cereals is a global food safety concern. One of the most common mycotoxins in grains is deoxynivalenol (DON), a secondary metabolite produced by the fungiFusarium graminearum and Fusarium culmorum. Exposure to DON can lead to adverse health effects in both humans and animals including vomiting, dizziness, and fever. Hence, the development of analytical technologies capable of predicting mycotoxin contamination levels in grains is crucial. In this study, we emphasize innovative infrared (IR) spectroscopic technologies for the prediction of DON in wheat along the food supply chain. The performance of an IR laser spectroscopic platform for on-site or laboratory confirmative analysis was evaluated. Furthermore, the performance of a handheld IR spectrometer for preliminary screening during transportation, storage, or harvesting was assessed. The accuracy of cross validation (AccCV) obtained with the laser spectrometer reached 92%, while the handheld IR spectrometer achieved 84.6%. Hence, both technologies prove significant potential for rapid mycotoxin detection.
- 8 January 2025
- Published by American Chemical Society (ACS) in ACS Food Science & Technology
- Vol. 5 (1) , 209-217
Open AccessInnovative Infrared Spectroscopic Technologies for the Prediction of Deoxynivalenol in Wheat
Mycotoxin contamination in cereals is a global food safety concern. One of the most common mycotoxins in grains is deoxynivalenol (DON), a secondary metabolite produced by the fungiFusarium graminearum and Fusarium culmorum. Exposure to DON can lead to adverse health effects in both humans and animals including vomiting, dizziness, and fever. Hence, the development of analytical technologies capable of predicting mycotoxin contamination levels in grains is crucial. In this study, we emphasize innovative infrared (IR) spectroscopic technologies for the prediction of DON in wheat along the food supply chain. The performance of an IR laser spectroscopic platform for on-site or laboratory confirmative analysis was evaluated. Furthermore, the performance of a handheld IR spectrometer for preliminary screening during transportation, storage, or harvesting was assessed. The accuracy of cross validation (AccCV) obtained with the laser spectrometer reached 92%, while the handheld IR spectrometer achieved 84.6%. Hence, both technologies prove significant potential for rapid mycotoxin detection.