Although natural anthocyanins (ACNs) possess strong antioxidant activity and pH-responsive properties, their development in food packaging films has been limited due to their instability and poor compatibility with the film matrix. This study developed a plasma-modified corn starch/polyvinyl alcohol (CS/PVA)-based film incorporating an anthocyanin-loaded ovalbumin-sodium alginate (OVA-SA) nanocomposite for salmon preservation and real-time freshness monitoring. Cold plasma treatment enhanced starch solubility and compatibility. The triple-crosslinking network achieved 81.94% ACNs encapsulation efficiency through ACNs-SA-OVA electrostatic/hydrogen bonding/it-it stacking interactions, OVA-SA electrostatic/hydrogen bonding interactions, and Ca2+-mediated crosslinking. The CP-OSA0.4 films demonstrated the highest mechanical properties (tensile strength: 30.51 MPa, elongation at break: 310.04%), water vapor barrier performance (6.28 & times; 10- 11 g & sdot;m- 1 & sdot;s- 1 & sdot;Pa- 1), and hydrophobicity (95.25 degrees). Rapid colorimetric response was observed upon exposure to ammonia, and the films exhibited reversible color changes from yellow-green to red, indicating their potential as freshness indicators. Additionally, the CP-OSA0.4 film exhibited strong antioxidant activity (DPPH and ABTS radical scavenging activities of 61.87% and 47.92%, respectively), excellent antibacterial activity, good biocompatibility, and complete biodegradation within 21 days, offering a promising eco-friendly alternative to conventional plastic packaging. In the salmon preservation trials, a strong correlation (R2 = 0.989) was observed between color changes and total volatile basic nitrogen levels. These films extended the shelf life of salmon by 4 days while providing visual freshness indication. Overall, this work provides a promising strategy for stabilizing ACNs and developing biodegradable intelligent packaging films for the real-time monitoring and maintenance of fish freshness.

Accurate authentication of rice geographical origin is crucial for food safety and fraud prevention. Synchrotron radiation X-ray fluorescence (SR-XRF) spectroscopy was combined with deep learning to classify hulled rice (n = 903) and rice husks (n = 824) from 16 provinces in China. Distinct elemental fingerprints were observed, and PCA/t-SNE visualization confirmed clustering by origin. However, PCA-based discrimination was limited in resolving provinces with overlapping profiles, necessitating advanced nonlinear approaches. Three deep learning models-1D-CNN, 2D-VGG16, and 2D-AlexNet-were trained on SR-XRF spectra. 2D-AlexNet achieved the best results, with accuracies of 98.02% for hulled rice and 99.07% for husks, showing strong robustness across provinces. Husk-based classification outperformed hulled rice, highlighting rice husks as a reliable sample matrix. The SR-XRF-deep learning framework provides a rapid, non-destructive, scalable tool for rice traceability, surpassing conventional chemometric or isotope-based methods.

Real-time monitoring of blueberry maturity is essential for automated harvesting and quality control. However, manual monitoring is labor-intensive and subjective, while mainstream visual models often possess redundant parameters that hinder deployment on low-power embedded devices. This study develops a low-cost, high-precision online monitoring system based on the Raspberry Pi platform, establishing a "perception-decision-feedback" closed-loop workflow. We propose a lightweight LIRE-YOLOv8 model employing a "front-end optimization, mid-end enhancement, and back-end efficiency" strategy. Key innovations include: (1) integrating a FasterNet-based backbone to reduce complexity; (2) implementing a dynamically weighted AFPN4 structure for asymptotic feature fusion; and (3) embedding the C2f-CBAM attention mechanism to enhance accuracy. The model size is compressed to 12.8MB (a 42% reduction), with a single-frame inference time of 0.32 seconds. Experimental results demonstrate that the system achieves an average recognition accuracy of 97.2% in real greenhouses. With a daily energy consumption of 5.04Wh, the system supports 53.8 days of continuous operation without external power, while hardware costs remain below $140.26. These findings validate its robustness and sustainability in complex field conditions, addressing the technical challenge of balancing high precision with low cost and power consumption. This research provides a practical blueprint for intelligent orchard management and the democratization of smart agriculture technologies.

The environmental risk assessment of CeO₂ engineered nanoparticles (ENPs) is limited by the inability to distinguish their specific exposure from ubiquitous natural counterparts (NNPs) in agroecosystems. Here, we developed an analytical framework integrating single-particle inductively coupled plasma time-of-flight mass spectrometry (spICP-TOFMS) with machine learning (ML) to enable source discrimination of Ce-containing particles in a soil-plant system. The method successfully differentiated CeO₂ ENPs from NNPs in both soil and maize (Zea mays L.) tissues against a natural Ce background (67.4 mg/kg). Source-resolved revealed that maize roots accumulate CeO₂ ENPs in a dose-dependent manner, with particle numbers increasing by one to two orders of magnitude relative to the control, whereas NNP uptake remained stable. In contrast, ENP levels in shoots were significantly lower, indicating limited upward translocation. These results demonstrate that separating engineered from natural nanoparticle fractions can improve exposure assessment beyond total element measurements. The proposed framework provides a practical approach for generating source-resolved exposure data and supports improved environmental risk assessment of engineered nanomaterials, although further standardization, validation, and sensitivity improvements are required for broader application.

Carbohydrate-polymer-based colorimetric films offer a sustainable platform for food freshness monitoring, but natural indicators often suffer from poor stability and premature migration. Here, a pH-responsive colorimetric indicator film was developed by loading anthocyanins (ACNs) into γ-cyclodextrin metal–organic frameworks (γ-CD-MOFs), coating them with an ovalbumin–sodium alginate (OVA-SA) shell, and embedding the nanocomposites in a plasma-modified corn starch/polyvinyl alcohol (PCS-PVA) matrix. The γ-CD-MOFs served as porous hosts for ACN confinement, while the OVA-SA shell helped regulate pigment release and access of alkaline volatiles. Molecular modeling and pore analysis suggested that ACNs could be accommodated within γ-CD-MOF cavities through size matching and noncovalent interactions. The resulting film showed 74.8% DPPH radical-scavenging activity, reduced ACN release, biodegradability, cytocompatibility, and clear pH/ammonia responsiveness. During salmon storage, film color evolution correlated strongly with total volatile basic nitrogen. EfficientNet-B0 achieved 96.7% classification accuracy under controlled laboratory conditions, and its generalizability to real-world applications remains to be externally validated. The structure–property results indicated that the carbohydrate-derived host framework and polysaccharide-containing shell jointly improved pigment confinement, interfacial compatibility, and colorimetric sensing performance. This work provides a carbohydrate-polymer strategy for stabilizing natural pigments and constructing freshness indicator films.

The quality evaluation of Zhenjiang aromatic vinegar (ZAV) is of paramount significance for standardizing the market and safeguarding consumers' interests. Near-infrared (NIR) spectroscopy combined with chemometrics was investigated to achieve the simultaneous discrimination of ZAV quality grades and quantification of key physicochemical parameters (total acid (TA), reducing sugar (RS), amino acid nitrogen (AAN)), thereby developing a rapid and comprehensive strategy of quality evaluation of ZAV. The denoising performance of five preprocessing methods was systematically evaluated prior to implementing competitive adaptive reweighted sampling (CARS) and variable combination population analysis (VCPA) algorithms for characteristic wavelength selection. The VCPA-based linear discriminant analysis model achieved a 97.14 % recognition rate of prediction set in ZAV grade discrimination. While partial least squares regression and support vector machine regression models based on the wavelengths selected by CARS yielded superior predictive ability for the three key parameters, with relative prediction deviation values of 4.03 (TA), 3.91 (RS) and 4.30 (AAN), respectively. This study provides a robust, cost-effective, and non-destructive alternative to traditional methods, establishing NIR spectroscopy as a highly promising tool for ensuring quality control and compliance in the vinegar industry.

Zearalenone (ZEN) is an estrogenic mycotoxin, posing a serious threat to food safety and human health. In this study, a new technique coupling ultraviolet (UV) and atmospheric cold plasma (ACP) was used to investigate how to degrade ZEN efficiently and to reveal the accelerated degradation mechanism. Within 1 min, UV and ACP together degraded 61.65 % of ZEN, which was greater than the sum of UV alone (16.67 %) and ACP alone (3.28 %), showing a synergistic effect. Four degradation products were identified using isotope tracing, LC-MS/MS, and NMR, namely UV-mediated isomerization product cis-ZEN and ACP-mediated oxidation products. After 10 min, ZEN was reduced from 0.58 μg to 0.08 μg. cis-ZEN accounted for more than 90 % among all products. The degradation acceleration was attributed to fast electron transitions and high energy release under high-voltage electric fields and ACP. This study provides a potential detoxification method for other mycotoxins, showing a broad application prospect.

Biocontrol is an effective technology for managing mycotoxin contamination in food, and the improvement of its application depends largely on revealing the degradation mechanisms at the molecular level. Research in this area is much less than that on the screening of degrading strains. In a previous study, Meyerozyma guilliermondii AF01 was confirmed to exert degradation and adsorption effects on aflatoxin B1 (AFB1). In this study, a potential degradation gene, MG2-4, was mined using a combination of bioinformatics and chemical approaches. The gene was heterologously expressed in Escherichia coli Rosetta DE3, and the recombinant protein, Mg aldo–keto reductase (AKR), reacted with AFB1 in vitro. Moreover, MgAKR rapidly removed AFB1. The degradation product was identified as aflatoxicol using ultra-high-performance liquid chromatography-quadrupole time-of-flight mass spectrometry, which is the same as the degradation product of the AF01 strain. This study reveals that MG2-4 is the key AFB1-degrading enzyme gene in the AF01 strain and lays the foundation for improving AFB1 removal using the AF01 strain.

Camellia oil (CAO), a unique edible plant oil native to China with distinct nutritional benefits, is often adulterated with cheaper oils due to its high profit. Therefore, the detection of the types and concentrations of these adulterants is crucial. In this study, the near-infrared (NIR) spectroscopy combined with two dimensional correlation spectroscopy (2DCOS) analysis method is proposed, revealing the interaction mechanism of adulterants in CAO, while enabling the rapid identification and quantification of adulteration in CAO. Firstly, Adulterated CAO samples with different concentration gradients were prepared, and the spectral characterization of the oil samples was completed. Then, the 2DCOS method was applied to distinguish highly overlapping spectral bands in adulterated oils of different concentrations. Synchronous spectra were employed to identify characteristic wavelengths from autocorrelation peaks, and asynchronous spectra were employed to analyze the sequential interaction of fatty acid fusion in adulterated oils under varying concentration gradients. Furthermore, the simplified multispectral models using support vector machine (SVM) and Partial Least Squares Regression (PLSR) were established with selected characteristic wavelengths as input variables. The established multispectral models demonstrated good performance with an accuracy of 92.78 %, the determination coefficients between 0.9601 and 0.9940, and the limits of detection between 0.70 % and 1.84 % for these adulterated oils. These results indicate that NIR spectroscopy combined with 2DCOS can be used to serve as a powerful rapid detection technique for adulteration analysis in CAO, and provides new methods for the interaction mechanism of adulterated oil components.

Color is an important quality indicator for agricultural products.The internal color of potatoes directly affects the sensory quality of their processed products.The rapid and non-destructive digital detection of the internal color of potatoes,as well as the quick classification of yellow-core and white-core potatoes,is of significant importance for advancing China's staple potato industry and the rapid rise of the prepared food industry.This study aims to achieve rapid,non-destructive digital characterization of the internal color of potatoes.It utilizes two self-developed,portable,multi-quality non-destructive detection devices based on a mini-spectrometer and a discrete spectral sensor:one is a laboratory-based,portable multi-quality non-destructive detection device for potatoes,and the other is a handheld,multi-quality non-destructive detection device for potatoes.A total of 209 potato samples from 26 different varieties,grown in various regions,were selected.Continuous spectral and discrete spectral data were collected,and the internal color parameters L*,a*and b*of the potato samples were measured using a colorimeter.First,based on the L*,a*and b*values measured by the colorimeter,an SVM classification model was established to distinguish between yellow-core and white-core potatoes.The threshold plane for distinguishing between yellow-and white-core potatoes was determined,which will serve as the basis for the next step in the non-destructive and rapid identification of yellow-and white-core potatoes.Secondly,based on the continuous spectral data collected by the portable continuous spectral device from the 209 potato samples,SNV preprocessing combined with the Random Frog Jump(RF)algorithm was used to select 200 characteristic wavelengths to establish a PLSR(Partial Least Squares Regression)prediction model for the L*,a*and b*parameters of potatoes.The root mean square errors(RMSE)for the validation set of L*,a*and b*were 1.278 8,0.081 6,and 1.407 1,respectively.Similarly,for the discrete spectral data collected by the handheld spectral device,after SNV preprocessing,PLSR prediction models for the L*,a*and b*parameters of potatoes were also established.The RMSE for the validation set of L*,a*and b*were 1.278 8,0.081 6,and 1.407 1,respectively.The results showed that the prediction models for the internal color parameters of potatoes established with both devices can meet the demand for rapid,non-destructive,digital detection of potato internal color in the field.Finally,52 potato samples from 26 varieties,which were not involved in model training,were selected for external validation of the L*,a*and b*predicted values using both devices.For the portable continuous spectral device,the maximum absolute residuals for L*,a*and b*were 2.617 3,0.141 3,and 2.779 1,respectively,and the mean residuals were 0.857 7,0.049 0,and 0.697 2,respectively.For the handheld discrete spectral device,the maximum absolute residuals for L*,a*and b*were 3.262 8,0.203 4,and 3.519 5,respectively,and the mean residuals were 1.093 0,0.066 7,and 1.268 8,respectively.Based on the L*,a*and b*predicted values from both devices,rapid non-destructive identification of yellow and white-core potatoes was performed using the previously established classification threshold plane.The classification accuracy for yellow and white-core potatoes was 92.31％for the portable continuous spectral device and 86.54％for the handheld discrete spectral device.This technology enables the rapid,non-destructive,and real-time digital detection of potato internal color,as well as the quick classification of yellow-and white-core potatoes in the field.It provides technical support for the entire potato industry chain,including planting,processing,and sales.

Near infrared spectroscopy (NIRS) has been widely used as a nondestructive testing technique and plays a crucial role in the quality inspection of agricultural products. However, the variability between different batches of samples hinders the application of commercial NIRS processes. Therefore, model transfer is usually performed on new samples to enhance the generalizability of the device. In this study, a new algorithm was developed based on the slope and bias correction algorithm (SBC) for model transfer between two different batches of samples, using potatoes of different origins as experimental samples for prediction models of potato quality, and based on this algorithm, model transfer between two different batches of samples was successfully implemented in a self-made portable non-destructive potato detection device. The results showed that the device developed based on the new algorithm gives good results for subsamples prediction. In the dry matter model, the correlation coefficient (R), root mean square error (RMSE) and relative standard deviation (RSD) of the new algorithm optimized compared with the traditional SBC algorithm were improved from 0.7843, 1.2080% and 6.59% to 0.8251, 1.1307 and 6.17%, respectively; in the starch model, the new algorithm optimized R, RMSE and RPD improved from 0.7971, 1.0023% and 7.43% to 0.8176, 0.9570% and 7.31%, respectively, compared with the traditional SBC algorithm. The transfer of NIR correction models for the dry matter and starch content of potatoes was basically achieved, which provided technical and theoretical support to enhance the model universality of convenient nondestructive detection devices.

IntroductionAccurate differentiation of benign and malignant pulmonary nodules in ultrasound remains a clinical challenge due to insufficient diagnostic precision. We propose the Deep Cross-Entropy Fusion (DCEF) model to enhance classification accuracy.MethodsA retrospective dataset of 135 patients (27 benign, 68 malignant training; 11 benign, 29 malignant testing) was analyzed. Manually annotated ultrasound ROIs were preprocessed and input into DCEF, which integrates ResNet, DenseNet, VGG, and InceptionV3 via entropy-based fusion. Performance was evaluated using AUC, accuracy, sensitivity, specificity, precision, and F1-score.ResultsDCEF achieved an AUC of 0.873 (training) and 0.792 (testing), outperforming traditional methods. Test metrics included 71.5% accuracy, 70.69% sensitivity, 70.58% specificity, 72.55% precision, and 71.13% F1-score, demonstrating robust diagnostic capability.DiscussionDCEF’s multi-architecture fusion enhances diagnostic reliability for ultrasound-based nodule assessment. While promising, validation in larger multi-center cohorts is needed to address single-center data limitations. Future work will explore next-generation architectures and multi-modal integration.

BackgroundMercury, as a global heavy metal pollutant, poses a serious threat to human health. The toxicity of mercury depends on its chemical form. Distinguishing the forms of mercury in the environment is of great significance for mercury management and reducing human mercury exposure risks. ObjectiveTo establish a non-targeted metallomics method based on synchrotron radiation X-ray fluorescence (SRXRF) spectroscopy combined with machine learning to screen inorganic mercury (IHg) or methylmercury (MeHg) exposed rice plants. MethodsRice seeds were exposed to ultra-pure water (control group), 0.1 mg·L−1 IHg (IHg group) or MeHg (MeHg group) solutions, respectively. After germination, the seedlings were cultured for 21 d, and rice leaves were collected, dried, weighed, and pressed. The content of metallome in rice leaves was determined by SRXRF. Machine learning models including soft independent modeling cluster analysis (SIMCA), partial least squares discriminant analysis (PLS-DA), and logistic regression (LR) were used to classify the SRXRF full spectra of different groups and find the best model to distinguish rice exposed to IHg or MeHg. Besides, characteristic elements were selected as input parameters to optimize the model by improving computing speed and reducing model calculation. ResultsThe SRXRF spectral intensities of the control group, IHg group, and MeHg group were different, indicating that exposure to IHg and MeHg can interfere the homeostasis of metallome in rice leaves. The results of principal component analysis (PCA) of SRXRF spectra showed that the control group could be well distinguished from the mercury exposed groups, but the IHg group and the MeHg group were mostly overlapped. The accuracy rates of the three models (PLS-DA, SIMCA, and LR) were higher than 98% for the training set, higher than 95% for the validation set, and higher than 94% for the cross-validation set. Besides, the accuracy of the LR model was higher than that of the PLS-DA model and the SIMCA model. Furthermore, the accuracy was 92.05% when using characteristic elements K, Ca, Mn, Fe, and Zn selected by LR to distinguish the IHg group and the MeHg group. Compared with the full spectra model, although the prediction accuracy of the characteristic spectral model decreased, the input parameters of the model decreased by 99.51%, and precision, recall, and F1 score were above 84.48%, indicating that the model could distinguish rice exposed to different mercury forms. ConclusionNon-targeted metallomics method based on SRXRF and machine learning can be applied for high-throughput screening of rice exposed to different forms of mercury and thus decrease the risks of people being exposed to mercury.

Microplastics (MPs) and nanoplastics (NPs) are global pollutants with emerging concerns. Methods to predict and screen their toxicity are crucial. Elemental dyshomeostasis can be used to assess toxicity of environmental pollutants. Non-targeted metallomics, combining synchrotron radiation X-ray fluorescence (SRXRF) and machine learning, has successfully differentiated cancer patients from healthy individuals. The whole idea of this work is to screen the phytotoxicity of nano polyethylene terephthalate (nPET) and micro polyethylene terephthalate (mPET) through non-targeted metallomics with SRXRF and deep learning algorithms. Firstly, Seed germination, seedling growth, photosynthetic changes, and antioxidant activity were used to evaluate the toxicity of mPET and nPET. It was showed that nPET, at 10 mg/L, was more toxic to rice seedlings, inhibiting growth and impairing chlorophyll content, MDA content, and SOD activity compared to mPET. Then, rice seedling leaves exposed to nPET or mPET was examined with SRXRF, and the SRXRF data was differentiated with deep learning algorithms. It was showed that the one-dimensional convolutional neural network (1D-CNN) model achieved 98.99% accuracy without data preprocessing in screening mPET and nPET exposure. In all, non-targeted metallomics with SRXRF and 1D-CNN can effectively screen the exposure and phytotoxicity of nPET/mPET and potentially other emerging pollutants. Further research is needed to assess the phytotoxicity of different types of MPs/NPs using non-targeted metallomics.

The use of microorganisms to manage aflatoxin contamination is a gentle and effective approach. The aim of this study was to test the removal of AFB1 from AFB1-contaminated peanut meal by a strain of Meyerozyma guilliermondii AF01 screened by the authors and to optimize the conditions of the biocontrol. A regression model with the removal ratio of AFB1 as the response value was established by means of single-factor and response surface experiments. It was determined that the optimal conditions for the removal of AFB1 from peanut meal by AF01 were 75 h at 29 °C under the natural pH, with an inoculum of 5.5%; the removal ratio of AFB1 reached 69.31%. The results of simulating solid-state fermentation in production using shallow pans and fermentation bags showed that the removal ratio of AFB1 was 68.85% and 70.31% in the scaled-up experiments, respectively. This indicated that AF01 had strong adaptability to the environment with facultative anaerobic fermentation detoxification ability. The removal ratio of AFB1 showed a positive correlation with the growth of AF01, and there were no significant changes in the appearance and quality of the peanut meal after fermentation. This indicated that AF01 had the potential to be used in practical production.

Allantoin, as a functional constituent of yam, has an extremely important role in the medical and cosmetic fields. In this study, based on the Raman spectroscopy detection system constructed in the laboratory, the Raman spectra of the powder of allantoin standard and the surface-enhanced Raman spectra of the allantoin extract of fresh yam were analyzed, and the surface-enhanced Raman characteristic displacements of allantoin in raw yam were determined to be 644, 1027 and 1398 cm(-1). The effects of the adsorption time of allantoin and silver sol and the thickness of the yam on the intensity of Raman feature displacement were investigated, and a method was established to directly obtain the surface-enhanced Raman feature information of allantoin in fresh yam. Based on this method, the surface-enhanced Raman spectra of 32 raw yams were collected, and the Raman feature displacements of allantoin at 644, 1027 and 1398 cm(-1) were established by unary linear regression (ULR), multivariable linear regression (MLR), and partial least squares regression (PLSR). The results showed that the MLR model was the most effective, with the validation set coefficient of determination (R-V(2)) of 0.93 and the root mean square error of validation (RMSEV) of 0.35 mg/g. However, the allantoin feature shift was susceptible to the changes of solution polarity and substrate, which led to a certain shift of the feature shift affecting the accuracy of the detection, and the quantitative prediction model of PLSR using the full-waveband Raman spectroscopy would improve the model ' s Robustness. The random frog (RF)-PLSR quantitative prediction model of allantoin was established based on the RF algorithm to screen the feature variables, and the R-V(2) was increased to 0.96, and the RMSEV was reduced to 0.26 mg/g. The model was externally validated using ten raw yam samples which were not involved in the modeling, and the absolute value of maximum residual was 0.74 mg/g. The method could realize the rapid quantitative detection of allantoin content in raw fresh yam, and provided new ideas and technical references for the direct rapid quantitative detection of allantoin in agricultural products.

A non-destructive method for determining the color value of pelletized red peppers is crucial for pepper processing factories. This study aimed to investigate the potentiality of visible and fluorescence images for the determination of color value of pelletized red pepper. The imaging problem, caused by the cylindrical shape and irregular cross-sectional features of the pelletized red peppers, was reduced through the extraction of an approximate plane region. To integrate the information in the visible and fluorescence images, a baseline convolutional neural network (CNN) architecture was designed and low level, middle level, and high level fusion models (denoted as LL-CNN, ML-CNN, and HL-CNN, respectively) were developed upon the baseline CNN. The effects of input image size and color space were examined. According to the training result, CNN fusion models were developed using visible image in L*a*b* color space and fluorescence image in RGB color space using 56 x 56 input image size. Among the three types of CNN fusion models, the HL-CNN obtained the best performance, resulting in Rv2 of 0.828 and RMSEV of 0.351. This study suggests that the fusion of visible and fluorescence image through CNN is a practical approach to save testing time and replace traditional destructive method. The low cost and compact structure of the imaging systems can maintain the commercial appeal of pepper industry.

Aspergillus flavus and its toxic metabolites-aflatoxins infect and contaminate maize kernels, posing a threat to grain safety and human health. Due to the complexity of microbial growth and metabolic processes, dynamic mechanisms among fungal growth, nutrient depletion of maize kernels and aflatoxin production is still unclear. In this study, visible/near infrared (Vis/NIR) hyperspectral imaging (HSI) combined with the scanning electron microscope (SEM) was used to elucidate the critical organismal interaction at kernel (macro-) and microscopic levels. As kernel damage is the main entrance for fungal invasion, maize kernels with gradually aggravated damages from intact to pierced to halved kernels with A. flavus were cultured for 0-120 h. The spectral fingerprints of the A. flavus-maize kernel complex over time were analyzed with principal components analysis (PCA) of hyperspectral images, where the pseudo-color score maps and the loading plots of the first three PCs were used to investigate the dynamic process of fungal infection and to capture the subtle changes in the complex with different hardness of the maize matrix. The dynamic growth process of A. flavus and the interactions of fungus-maize complexes were explained on a microscopic level using SEM. Specifically, fungus morphology, e.g., hyphae, conidia, and conidiophore (stipe) was accurately captured on the microscopic level, and the interaction process between A. flavus and nutrient loss from the maize kernel tissues (i.e., embryo, and endosperm) was described. Furthermore, the growth stage discrimination models based on PLSDA with the results of CCRC = 100 %, CCRV = 97 %, CCRIV = 93 %, and the prediction models of AFB1 based on PLSR with satisfactory performance (R2C = 0.96, R2V = 0.95, R2IV = 0.93 and RPD = 3.58) were both achieved. In conclusion, the results from both macro-level (Vis/NIR-HSI) and micro-level (SEM) assessments revealed the dynamic organismal interactions in A. flavus-maize kernel complex, and the detailed data could be used for modeling, and quantitative prediction of aflatoxin, which would establish a theoretical foundation for the early detection of fungal or toxin contaminated grains to ensure food security.

Ambient air pollution is an important contributor to increasing cases of lung cancer, which is a malignant cancer with the highest mortality among all cancers. It primarily manifests in the form of pulmonary nodules, but not all will develop into lung cancer. Therefore, it is highly desired to distinguish between benign and malignant pulmonary nodules for the early prevention and treatment of lung cancer. Currently, histopathological examination is the gold standard for classifying pulmonary nodules, which is invasive, time-consuming, and labor-intensive. This study proposes a metallomics approach through synchrotron radiation X-ray fluorescence (SRXRF) with a simplified one-dimensional convolutional neural network (1DCNN) to distinguish pulmonary nodules by using serum samples. SRXRF spectra of serum samples were obtained and preliminarily analyzed using principal component analysis (PCA). Subsequently, machine learning algorithms (MLs) and 1DCNN were applied to develop classification models. Both MLs and 1DCNN based on full-channel spectra could distinguish patients with benign and malignant pulmonary nodules, but the highest accuracy rate of 96.7% was achieved when using 1DCNN. In addition, it was found that characteristic elements in serum from patients with malignant nodules were different from those in benign nodules, which can serve as the fingerprint metallome profile. The simplified model based on characteristic elements resulted in good performance of sensitivity and F1-score > 91.30%, G-mean, MCC and Kappa > 85.59%, and accuracy = 94.34%. In summary, metallomic classification of benign and malignant pulmonary nodules using serum samples can be achieved through 1DCNN-boosted SRXRF, which is easy to handle and much less invasive compared to histopathological examination.

Maturity is a crucial indicator in assessing the quality of tomatoes, and it is closely related to lycopene content. Using hyperspectral imaging, this study aimed to monitor tomato maturity and predict its lycopene content at different maturity stages. Standard normal variable (SNV) transformation was applied to preprocess the hyperspectral data. Then, using competitive adaptive reweighted sampling (CARS), the characteristic wavelengths were selected to simplify the calibration models. Based on the full and characteristic wavelengths, a support vector classifier (SVC) model was developed to determine tomato maturity qualitatively. The results demonstrated that the classification accuracy using the characteristic wavelength led to the obtention of better results with an accuracy of 95.83%. In addition, the support vector regression (SVR) and partial least squares regression (PLSR) models were utilized to predict lycopene content. With a coefficient of determination for prediction (R2P) of 0.9652 and a root mean square error for prediction (RMSEP) of 0.0166 mg/kg, the SVR model exhibited the best quantitative prediction capacity based on the characteristic wavelengths. Following this, a visual distribution map was created to evaluate the lycopene content in tomato fruit intuitively. The results demonstrated the viability of hyperspectral imaging for detecting tomato maturity and quantitatively predicting the lycopene content during storage.