Time series prediction of insect pests in tea gardens.

BACKGROUND: Tea-garden pest control is crucial to ensure tea quality. In this context, the time-series prediction of insect pests in tea gardens is very important. Deep-learning-based time-series prediction techniques are advancing rapidly but research into their use in tea-garden pest prediction is limited. The current study investigates the time-series prediction of whitefly populations in the Tea Expo Garden, Jurong City, Jiangsu Province, China, employing three deep-learning algorithms, namely Informer, the Long Short-Term Memory (LSTM) network, and LSTM-Attention. RESULTS: The comparative analysis of the three deep-learning algorithms revealed optimal results for LSTM-Attention, with an average root mean square error (RMSE) of 2.84 and average mean absolute error (MAE) of 2.52 for 7 days' prediction length, respectively. For a prediction length of 3 days, LSTM achieved the best performance, with an average RMSE of 2.60 and an average MAE of 2.24. CONCLUSION: These findings suggest that different prediction lengths influence model performance in tea garden pest time series prediction. Deep learning could be applied satisfactorily to predict time series of insect pests in tea gardens based on LSTM-Attention. Thus, this study provides a theoretical basis for the research on the time series of pest and disease infestations in tea plants. © 2024 Society of Chemical Industry.

Rapid and stable detection of three main mycotoxins in rice using SERS optimized AgNPs@K30 coupled multivariate calibration.

Food safety is regarded as a crucial factor in both human health and economic progress. This study focuses on the fabrication of a label-free surface-enhanced Raman scattering (SERS) sensor for rapid sensing of three main mycotoxins (aflatoxin B1 (AFB1), ochratoxin A (OTA), ochratoxin B (OTB)) in rice using the optimized rough silver nanoparticles (AgNPs@K30) with enhancement factor (EF) 1.58 × 107 coupled multivariate calibration. Two variable selection chemometric calibration methods were comparatively applied. And genetic algorithm-partial least square achieved optimum correlation coefficient = 0.9797, 0.9779, and 0.9827, respectively for AFB1 ranging from 0.5 to 250 µg/Kg, for OTA and OTB 1 to 500 µg/Kg. The limit of detection (LOD) = 1.145, 1.133, and 1.180 µg/Kg, respectively, were determined according to principal component analysis-calibrated loading weight approach. And the great stability and reproducibility revealed the prepared SERS sensor has the potential to predict AFB1, OTA, and OTB in real rice samples.

Simultaneous determination of benzimidazole fungicides in food using signal optimized label-free HAu/Ag NS-SERS sensor.

Extensively employed pesticide in agriculture causes residue in food products that would threaten public health safety. The surface-enhanced Raman scattering (SERS) signal reliant on double sensing of carbendazim and thiabendazole in a single step is achieved without the aid of any bio-recognition element. A label-free anisotropic bimetallic hollow Au/Ag nanostars (HAu/Ag NS) SERS substrate was synthesized with numerous hot spots for Raman molecule through a galvanic displacement-free deposition. The individual and mixed analyte calibration results were compared based on the identified peak at 1224 (carbendazim) and 778 (thiabendazole) cm-1 and exhibited insignificant differences. The sensor could detect carbendazim and thiabendazole up to 4.28 × 10-4 and 6.04 × 10-4 µg·g-1 or µg·mL-1 in both individual and mixture of their extract. The recovery for accuracy and precision analysis was 91.54-98.26 % in rice and water. Finally, validation results were achieved satisfactorily (p > 0.05) with HPLC.

Catalytic hairpin activated gold-magnetic/gold-core-silver-shell rapid self-assembly for ultrasensitive Staphylococcus aureus sensing via PDMS-based SERS platform.

Staphylococcus aureus (S. aureus) has been identified as a marker of food contamination, closely associated with human health. This work designs a sensitive and rapid bio-detection strategy for S. aureus based on hybridization chain reaction-assisted surface enhanced Raman scattering (HCR-assisted-SERS) signal amplification. In this approach, the interaction between the aptamer (Apt) and its partial complementary DNA strands (cDNA) fabricated on the surface of gold-assisted magnetic nanoparticles (Au-MNPs) and the subsequent detachment of the cDNA results in the activation of the HCR process. In the HCR, a pair of hairpin structured DNA probes (H1 and H2) with sticky ends self-assembles to form a long DNA polymer. Subsequently, the output and amplification of the SERS signal were performed by conjugating 4-ATP modified Au@Ag NPs with the obtained DNA polymer via a specific Ag-S bond, and further collected through a self-administered polydimethylsiloxane (PDMS) cone-shaped support array. The precise quantification of S. aureus was performed in the concentration range of 28 to 2.8 × 106 cfu/mL, achieving a detection limit of 0.25 cfu/mL. This strategy was further applied to S. aureus detection in spiked milk samples with good recoveries (91-102%) and the relative standard deviation (4.35-8.41%). The sensing platform also showed satisfactory validation results (p > 0.05) using the traditional plate counting method. The proposed HCR-assisted SERS probe can be extended to other foodborne pathogenic bacteria types via engineering appropriate Apt and DNA initiators, thus, inspiring widespread applications in food safety and biomedical research.

Rapid detection and prediction of chloramphenicol in food employing label-free HAu/Ag NFs-SERS sensor coupled multivariate calibration.

Considering growing food safety issues, hollow Au/Ag nano-flower (HAu/Ag NFs) nanosensor has been synthesized for label-free and ultrasensitive detection of chloramphenicol (CP) via integrating the surface-enhanced Raman scattering (SERS) and multivariate calibration. As the anisotropic plasmonic nanomaterials, HAu/Ag NFs had numerous nano-chink on their surface, which offered huge hotspots for analytes. CP generated a strong SERS signal while adsorbed on the surface of HAu/Ag NFs and noted excellent linearity with 1st derivative-competitive adaptive reweighted sampling-partial least squares (CARS-PLS) in the range of 0.0001-1000 µg/mL among the four applied multivariate calibrations. Additionally, CARS-PLS generated the lowest prediction error (RMSEP) of 0.089 and 0.123 µg/mL for milk and water samples, respectively, and any CARS-PLS model could be used for both samples according to T-test results (P > 0.05). The intra- and interday recovery for both samples were in the range of 92.62-96.74% with CV < 10%, suggested the proposed method has excellent accuracy and precision.

Rapid detection of mercury in food via rhodamine 6G signal using surface-enhanced Raman scattering coupled multivariate calibration.

Considering food safety and limitations of biorecognition elements, this study focused on the development of a novel method for predicting mercury (Hg2+) in fish and water samples using surface-enhanced Raman scattering (SERS) coupled wavenumber selection chemometric method. Herein, core-shell Au@Ag nanoparticles (Au@Ag NPs) were synthesized as SERS substrate, and rhodamine 6G (R6G) was used as signaling probe for Hg2+. In the presence of Hg2+, citrate ion of Au@Ag NPs induced complexation and become amalgam causes desorption of R6G occurred, resulted in decreased SERS signal intensity. Compared to surface Plasmon resonance method, SERS coupled genetic algorithm-partial least squares realized good correlation coefficient (0.9745 and 0.9773) in their prediction over the concentration ranges 1.0 × 102 to 1.0 × 10-3 µg/g. The recovery (88.45 - 94.73%) and precision (coefficient of variations, 3.28 - 5.76%) exhibiting satisfactory results suggested that the proposed method could be employed to predict Hg2+ in fish and water samples towards quality and safety monitoring.

Rapid measurement of fatty acid content during flour storage using a color-sensitive gas sensor array: Comparing the effects of swarm intelligence optimization algorithms on sensor features.

The fatty acid content of flour is an important indicator for determining the deterioration of flour. We propose a novel rapid measurement method for fatty acid content during flour storage based on a self-designed color-sensitive gas sensor array. First, a color-sensitive gas sensor array was prepared to capture the odor changes during flour storage. The sensor features were then optimized using genetic algorithm (GA), ant colony optimization (ACO) and particle swarm optimization (PSO). Finally, back propagation neural network (BPNN) models were established to measure the fatty acid content during flour storage. Experimental results showed that the optimization effects of the three algorithms improved in the following order: GA < ACO < PSO, for the sensor features optimization. In the validation set, the determination coefficient of the best PSO-BPNN model was 0.9837. The overall results demonstrate that the models established on the optimized features can realize rapid measurements of fatty acid content during flour storage.

Application of volatile and spectral profiling together with multimode data fusion strategy for the discrimination of preserved eggs.

The maturity level of eggs during pickling is conventionally assessed by choosing few eggs from each curing batch to crack open. Yet, this method is destructive, creates waste and has consequences for financial losses. In this work, the feasibility of integrating electronic nose (EN) with reflectance hyperspectral (RH) and transmittance hyperspectral (TH) data for accurate classification of preserved eggs (PEs) at different maturation periods was investigated. Classifier models based solely on RH and TH with EN achieved a training accuracy (93.33%, 97.78%) and prediction accuracy (88.89%; 93.33%) respectively. The fusion of the three datasets, (EN + RH + TH) as a single classifier model yielded an overall training accuracy of 98.89% and prediction accuracy of 95.56%. Also, 52 volatile compounds were obtained from the PE headspace, of which 32 belonged to seven functional groups. This study demonstrates the ability to integrate EN with RH and TH data to effectively identify PEs during processing.

Evaluation of matcha tea quality index using portable NIR spectroscopy coupled with chemometric algorithms.

BACKGROUND: The study reports a portable near infrared (NIR) spectroscopy system coupled with chemometric algorithms for prediction of tea polyphenols and amino acids in order to index matcha tea quality. RESULTS: Spectral data were preprocessed by standard normal variate (SNV), mean center (MC) and first-order derivative (1st D) tests. The data were then subjected to full spectral partial least squares (PLS) and four variable selection algorithms, such as random frog partial least square (RF-PLS), synergy interval partial least square (Si-PLS), genetic algorithm-partial least square (GA-PLS) and competitive adaptive reweighted sampling partial least square (CARS-PLS). RF-PLS was established and identified as the optimum model based on the values of the correlation coefficients of prediction (RP ), root mean square error of prediction (RMSEP) and residual predictive deviation (RPD), which were 0.8625, 0.82% and 2.13, and 0.9662, 0.14% and 3.83, respectively, for tea polyphenols and amino acids. The content range of tea polyphenols and amino acids in matcha tea samples was 8.51-14.58% and 2.10-3.75%, respectively. The quality of matcha tea was successfully classified with an accuracy rate of 83.33% as qualified, unqualified and excellent grade. CONCLUSION: The proposed method can be used as a rapid, accurate and non-destructive platform to classify various matcha tea samples based on the ratio of tea polyphenols to amino acids. © 2019 Society of Chemical Industry.