Aquaphotomic NIR Spectroscopy Technique Could Rapidly Detect Toxic Aflatoxin in Maize

Introduction

Maize is one of the world’s most widely consumed staple crops, providing a significant portion of calories and protein in global diets. However, contamination by aflatoxin-producing fungi such as Aspergillus flavus and Aspergillus parasiticus poses a serious risk to food safety. Aflatoxins are carcinogenic, and contamination can exceed regulatory limits in the European Union and the U.S., particularly in regions where monitoring is limited. Conventional detection methods, including enzyme-linked immunosorbent assay (ELISA), chromatographic techniques, and electrochemical immunosensors, are reliable but time-consuming and costly (1).

In a recent study, Appaw and colleagues explored the use of visible and near-infrared spectroscopy (Vis/NIRS) combined with chemometric and aquaphotomic methods to classify and quantify aflatoxin levels in maize (1). This approach offers a rapid, non-destructive alternative capable of supporting real-time, on-site decisions in the food supply chain (1).

Study Details and Methods

The research team, comprising W. Appaw, J.L.Z. Zaukuu, B. Aouadi, E.T. Mensah, I.N. Oduro, and Z. Kovacs, evaluated white and yellow maize samples with varying aflatoxin concentrations (0, 3, 5, 10, 20, 30, and 50 ng/g). The study used three datasets: naturally contaminated white maize, spiked white maize, and spiked yellow maize. Spectral data were collected across multiple wavelength ranges (450–1050 nm, 1150–2400 nm, and 1300–1600 nm) to capture relevant absorption peaks associated with aflatoxin presence. Key absorption bands were observed at 500, 950, 1000, 1300, 1500, 1900, 2100, and 2300 nm (1).

To improve model accuracy, the team applied Savitzky-Golay smoothing (first derivative, filter 17) during preprocessing. Principal component analysis combined with linear discriminant analysis (PCA-LDA) enabled classification of contaminated samples. Partial least squares regression (PLSR) was applied for quantitative prediction of aflatoxin concentrations (1).

Spectroscopic Findings

The study revealed that PCA-LDA models achieved up to 92.5% classification accuracy across combined datasets and 100% accuracy when models were developed separately for each maize type. Sensitivity, specificity, precision, and F1 scores were close to 1, reflecting robust classification (1).

PLSR models demonstrated strong predictive performance, with R²CV of 0.99, RMSECV of 1.70 ng/g, RPD of 9.90, LOD of 0.60 ng/g, and LOQ of 1.81 ng/g at 450–1050 nm. Similar results were obtained for higher wavelength ranges, underscoring the reliability of Vis-NIRS for aflatoxin quantification (1).

Aquaphotomics Insights

The researchers incorporated aquaphotomics to investigate water’s role as a molecular marker for contamination. By analyzing water matrix coordinates (WAMACs), the team visualized changes in water spectral patterns through aquagrams, which correlated with aflatoxin presence. These findings suggest that water molecules within maize can serve as sensitive indicators of contamination, offering an additional layer of analytical insight beyond conventional spectroscopy (1).

How Aquaphotonics Was Used

In this study, aquaphotonics was applied to investigate how water molecules within maize respond to the presence of aflatoxin, providing a sensitive marker for contamination. By analyzing NIR spectra, the researchers identified specific WAMACs that indicate changes in the molecular structure and interactions of water in the samples (1–3). These WAMACs were visualized using aquagrams, star-chart representations of the water spectral pattern, allowing the team to observe shifts associated with different levels of aflatoxin. The aquaphotomic approach complemented conventional chemometric methods, offering an additional, non-destructive layer of insight by linking changes in the water molecular network directly to contamination, suggesting that water spectral patterns can serve as reliable indicators of aflatoxin presence in both white and yellow maize (1–3).

Implications and Recommendations

The study highlights the potential for Vis/NIRS and aquaphotomics to streamline aflatoxin detection in maize. Researchers recommend developing region-specific calibration models, integrating cloud-based data storage for real-time supply chain monitoring, and conducting pilot programs in maize-producing regions to validate operational efficacy (1).

This non-destructive, rapid, and precise approach may improve food safety monitoring and reduce reliance on labor-intensive laboratory methods. By leveraging spectral data and aquaphotomic patterns, stakeholders can make faster, data-driven decisions to prevent contaminated maize from entering the food chain (1).

References
(1) Appaw, W.; Zaukuu, J. L. Z.; Aouadi, B.; Mensah, E. T.; Oduro, I. N.; Kovacs, Z. Predicting Aflatoxin Contamination in White and Yellow Maize Using Vis/NIR Spectroscopy Combined with PCA-LDA and PLSR Models through Aquaphotomics Approaches. Appl. Food Res. 2025, 5 (1), 100841. DOI: 10.1016/j.afres.2025.100841.

(2) Roger, J. M.; Mallet, A.; Marini, F. Preprocessing NIR Spectra for Aquaphotomics. Molecules 2022, 27 (20), 6795. DOI: 10.3390/molecules27206795.

(3) Tsenkova, R.; Munćan, J.; Pollner, B.; Kovacs, Z. Essentials of Aquaphotomics and Its Chemometrics Approaches. Front. Chem. 2018, 6, 363. DOI: 10.3389/fchem.2018.00363.