Microplastics (MPs) are emerging contaminants that pose potential health risks through bioaccumulation. Efficient detection of MPs in complex food matrices, such as high-value camel milk powder (CMP), is critical for food safety. This study investigated the feasibility of using Near-Infrared (NIRS) and Mid-Infrared Spectroscopy (MIRS) combined with data fusion strategy for the rapid qualitative and quantitative analysis of MPs in CMP. Samples with MP concentrations (PP, PE, PET, PS) ranging from 0.01% to 1.00% (w/w) were prepared. The PCA results demonstrated that NIRS and MIRS provide complementary information. NIRS primarily captures overtones and combination bands of fundamental molecular vibrations, whereas MIRS detects fundamental vibrational modes that offer specific structural fingerprints. For qualitative analysis, the PLS-DA model based on low-level fusion achieved the best performance with 100% classification accuracy. For quantitative analysis, mid-level fusion yielded the most robust PLS models, with R2CV for PE, PET, PP, and PS reaching 0.745, 0.855, 0.911, and 0.906, respectively. This study establishes a novel, non-destructive, and efficient approach for monitoring microplastic contamination in CMP.
Milk fat is a valuable ingredient in dairy products and is frequently subjected to adulteration with cheaper vegetable fats and oils. Butyric acid is a specific marker of dairy fats, while n-3 fatty acids are typically absent or present only in trace amounts in milk fat, whereas they may occur in significant concentrations in vegetable oils. Conventional methods for butyric acid determination, such as gas chromatography or Reichert–Meissl and Polenske indices, are laborious, time-consuming, destructive, or require hazardous reagents. Although NMR spectroscopy offers a rapid and non-destructive alternative, conventional 1H NMR fails to distinguish butyric and n-3 fatty acid moieties due to signal overlap of their terminal methyl groups. In this study, a rapid two-dimensional COSY (COrrelation SpectroscopY) NMR approach is proposed to distinguish butyric vs n-3 fatty acid moieties in fats and oils. The method exploits specific vicinal proton–proton correlations of the terminal methyl group, allowing discrimination between butyrate-specific A–D correlations and n-3 fatty acid–specific A–E correlations. COSY spectra of reference triglycerides, vegetable oils rich in n-3 fatty acids, dairy fats, and their admixtures clearly demonstrate the selectivity of this approach. The proposed COSY NMR method enables the identification of dairy, non-dairy, and mixed fat samples within approximately 15 min, without sample derivatization or destruction. This rapid, non-invasive technique represents a valuable screening tool for dairy fat authentication and adulteration detection.
Due to its high market demand and its exceptional nutritional value, pomegranate juice is an expensive product particularly vulnerable to adulteration. This fraud consists of diluting with cheaper juices, such as apple or grape juice, or masking poor-quality pomegranate juice with dark colored juices such as blueberry or black grape juice. To address this issue, a simple, rapid, and green methodology based on headspace gas chromatography-ion mobility spectrometry (HS-GC-IMS) combined with chemometrics was proposed, to detect adulteration using volatile organic compounds (VOCs) as markers. Spectra from 295 samples of pure and adulterated pomegranate juice were analyzed using a non-targeted approach and, in addition, 28 VOCs were identified.Orthogonal partial least squares discriminant analysis (OPLS-DA) models enabled the classification of samples as pure or adulterated pomegranate juice (100 % validation success rate) and discriminated among the three studied adulterants (blueberry, apple, and grape juices) with misclassifications occurring only at low levels of adulteration. Partial least squares (PLS) regression models were also developed to estimate the percentage of adulterants. High greenness (AGREE score of 0.73/1) and applicability (BAGI score of 72.5/100), together with encouraging analytical performance, highlight the proposed approach as a promising tool for pomegranate juice authentication.
Clostridium perfringens is a causative agent of food-borne disease due to its rapid growth and toxin production. Compared with conventional antimicrobials, antimicrobial peptides are less likely to induce bacterial resistance due to their multiple killing mechanisms. Hence, this study aimed to evaluate the antimicrobial characteristics and action modes of the AMP NKL-24 against C. perfringens vegetative cells and biofilms. The antimicrobial mechanism results revealed that NKL-24 could induce mitochondrial membrane depolarization and reduce intracellular ATP from 1074.8 to 935.8 nM as the concentration increased from 0 to 40 μg/mL. Assessment of membrane permeability using SYTO 9/Propidium iodide staining showed that 92.7% of the cells became PI-positive after treatment with 40 μg/mL NKL-24. Scanning electron microscopy images demonstrated that NKL-24 could cause cell disruption by compromising membrane integrity. Furthermore, admirable inhibitory and eradication effects of NKL-24 on C. perfringens biofilms were determined by crystal violet staining and confocal laser scanning microscopy analysis. Its antibiofilm activities involved inhibiting cellular viability and disrupting protein-polysaccharide matrix assembly in the biofilms. NKL-24 exhibited negligible cytotoxic effects on Caco-2 cells and no signs of acute toxicity in mice, along with normal serum and organ function profiles. Finally, pork sausage preservation demonstrated that the combination use of NKL-24 with sodium nitrite significantly extended the shelf life of the samples by suppressing microbial growth and lipid oxidation. These results highlight the potential of NKL-24 as a promising safe and multifunctional food preservative that can effectively reduce the required concentration of sodium nitrite in pork sausages.
Toxic Illicium adulterants are visually similar to Illicium verum (genuine star anise), making rapid, non-destructive authentication difficult in routine food safety inspection. This study developed a fruit-level hyperspectral discrimination framework based on a deep derivative-enhanced spectral attention network (DSAN) to distinguish Illicium verum from toxic Illicium adulterants (Illicium lanceolatum and Illicium henryi). Intact Illicium samples were imaged using visible/near-infrared (Vis/NIR) hyperspectral imaging, and fruit-level representative spectra were extracted from segmented plate-level hyperspectral cubes for subsequent analysis. A lightweight wavelength-wise spectral attention mechanism was applied to a derivative-enhanced input formed by concatenating raw spectra with their first- and second-order numerical derivatives, enabling the network to emphasize discriminative spectral features. The evaluation of model performance was conducted using plate-grouped leave-one-plate-out validation with 10 repeated runs. The results were then compared with four classical chemometric models and five deep learning baselines. To prevent the leakage of information, feature filtering and z-score normalization were performed using training-fold statistics. DSAN achieved an accuracy of 99.13 ± 0.41%, a balanced accuracy of 99.03 ± 0.43%, a recall of 98.54 ± 0.73% for toxic adulterants, and an F1-score of 98.90 ± 0.52%. In comparison, the baselines achieved 80.25-97.08% accuracy, 79.93-96.53% balanced accuracy, 74.69-96.88% recall, and 75.96-96.26% F1-score. These results demonstrate the effectiveness of derivative-enhanced spectral attention learning for fruit-level discrimination of genuine Illicium verum from toxic Illicium adulterants.
To address the low accuracy and efficiency of conventional single-modality approaches for discriminating moldy pear core, this study proposes a nondestructive detection method based on data-level fusion of vibration acoustic signals and Visible/Near-infrared (Vis/NIR) spectral signals to enhance the detection of moldy pear core. First, the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) was employed to reduce the signal errors caused by noise interference. Subsequently, the vibration acoustic signals and Vis/NIR spectral signals were fused and transformed into two-dimensional images using the continuous wavelet transform (CWT) algorithm. The texture features were extracted from the CWT images using the gray-level co-occurrence matrix–markov random field (GLCM-MRF). The feature extractor TransXNet was constructed to extract the deep features of CWT images. Finally, the texture features and deep features were extracted separately from the CWT images and evaluated as two alternative feature-extraction routes. The GLCM-MRF texture-feature route was used as a baseline comparison, whereas the final optimized model was constructed using TransXNet deep features followed by SHAP-based feature selection and XGBoost classification. The research results show that the TXN-XGBoost model constructed based on deep feature extractor performs the best in detection performance. The values of the weighted average precision (wAP), weighted average recall (wAR) and weighted average F1-score (wAF1) have all reached around 97 %. Furthermore, the SHapley Additive exPlanations (SHAP) analysis was employed to eliminate redundant features, and a TXN-SHAP-XGBoost model was developed based on the selected 158 positive features. Compared with the TXN + XGBoost model, the TXN-SHAP-XGBoost model improved the wAP, wAR, and wAF1 indicators of four types of fragrant pears to 98.28 %, 98.26 %, and 98.26 %. The discrimination time for a single pear was reduced to 0.13 s after SHAP-based feature selection. Therefore, this study provided an effective multimodal modeling strategy for accurate detection of moldy pear core and offered a useful basis for the future development of online nondestructive inspection systems for pears.
Tea, one of the most widely consumed non-alcoholic beverages worldwide, is increasingly challenged by contamination with anthraquinone (AQ), a polycyclic aromatic ketone with recognised genotoxic potential and the capacity to induce oxidative stress. AQ has been repeatedly detected in tea and tea-based products in recent years, and stringent maximum residue limits (MRLs) implemented in the European Union and other regions have made it a critical constraint on international tea trade and a potential risk to consumer health. This review summarizes the key sources and pathways of AQ contamination in tea, delineating an intricate contamination network involving long-range atmospheric transport and dry/wet deposition, agricultural inputs, processing- and storage-related formation and transfer, and potential endogenous biosynthesis. By integrating key toxicological evidence, including the association of AQ with perturbation of cytochrome P450 systems and DNA damage pathways, this study underscores the necessity for refined exposure characterisation and health risk assessment, particularly in populations with high tea consumption. Recent progress in instrumental methods for AQ determination in tea is also reviewed, with particular emphasis on gas chromatography–mass spectrometry (GC–MS) and related hyphenated techniques, and the analytical prospects of emerging tools, including molecularly imprinted polymers (MIPs) for selective enrichment and DNA-based biosensors for rapid screening, are discussed. Finally, an end-to-end quality control framework integrating source tracing, advanced detection, and sustainable processing management is proposed to support researchers, regulators, and the tea industry in reducing trade barriers, improving tea safety, and protecting public health.
