Our previous study showed that COPPER-CONTAINING AMINE OXIDASE (CuAO) and AMINOALDEHYDE DEHYDROGENASE (AMADH) could regulate the accumulation of γ-aminobutyric acid (GABA) in tea through the polyamine degradation pathway. However, their biological function in drought tolerance has not been determined. In this study, Camellia sinensis (Cs) CsCuAO1 associated with CsAMADH1 conferred drought tolerance, which modulated GABA levels in tea plants. The results showed that exogenous GABA spraying effectively alleviated the drought-induced physical damage. Arabidopsis lines overexpressing CsCuAO1 and CsAMADH1 exhibited enhanced resistance to drought, which promoted the synthesis of GABA and putrescine by stimulating reactive oxygen species’ scavenging capacity and stomatal movement. However, the suppression of CsCuAO1 or CsAMADH1 in tea plants resulted in increased sensitivity to drought treatment. Moreover, co-overexpressing plants increased GABA accumulation both in an Agrobacterium-mediated Nicotiana benthamiana transient assay and transgenic Arabidopsis plants. In addition, a GABA transporter gene, CsGAT1, was identified, whose expression was strongly correlated with GABA accumulation levels in different tissues under drought stress. Taken together, CsCuAO1 and CsAMADH1 were involved in the response to drought stress through a dynamic GABA-putrescine balance. Our data will contribute to the characterization of GABA’s biological functions in response to environmental stresses in plants.
Polyamines are a potential source of γ-aminobutyric acid (GABA) in plants under abiotic stress. However, studies on GABA enrichment in tea mostly focus on the GABA shunt, while the correlation between polyamine degradation and GABA formation in tea is largely unknown. In this study, tea plants responded to exogenous putrescine, resulting in a significant increase in GABA content, while the glutamate level did not change. At the same time, five copper-containing amine oxidase (CuAO) and eight aminoaldehyde dehydrogenase (AMADH) genes involved in the putrescine-derived GABA pathway were identified from the Tea Plant Information Archive. Expression analysis indicated that CsCuAO1, CsCuAO3 as well as CsAMADH1 were induced to play an important function in response to exogenous putrescine. Thus, the three genes were cloned and the catalytic efficiency of soluble recombinant proteins was determined. CsCuAOs and CsAMADH1 exhibited indispensable functions in the GABA production from putrescine in vitro. Subcellular localization assays indicated that CsAMADH1 was localized in plastid, while both CsCuAO1 and CsCuAO3 were localized in peroxisome. In addition, the synergistic effects of CsCuAOs and CsAMADH1 were investigated by a transient co-expression system in Nicotiana benthamiana. Our data suggest that these three genes regulate the accumulation of GABA in tea by participating in the polyamine degradation pathway and improve the content of GABA in tea to a certain extent. The results will greatly contribute to the production of GABA tea.
In recent years, convolutional‐neural‐network based stereo matching methods have achieved significant gains compared to conventional methods in terms of both speed and accuracy. Current state‐of‐the‐art disparity estimation algorithms require many parameters and large amounts of computational resources and are not suited for applications on edge devices. In this paper, an end‐to‐end light‐weight network (LWNet) for fast stereo matching is proposed, which consists of an efficient backbone with multi‐scale feature fusion for feature extraction, a 3D U‐Net aggregation architecture for disparity computation, and color guidance in a 2D convolutional neural network (CNN) for disparity refinement. MobileNetV2 is adopted as an efficient backbone in feature extraction. The channel attention module is applied to improve the representational capacity of features and multi‐resolution information is adaptively incorporated into the cost volume via cross‐scale connections. Further, a left‐right consistency check and color guidance refinement are introduced and a robust disparity refinement network is designed with skip connections and dilated convolutions to capture global context information and improve disparity estimation accuracy with little computational cost and memory space. Extensive experiments on Scene Flow, KITTI 2015, and KITTI 2012 demonstrate that the proposed LWNet achieves competitive accuracy and speed when compared with state‐of‐the‐art stereo matching methods.
Achieving rapid and non-destructive assessment of tea quality is essential for intelligent tea production and quality control. In this study, an integrated hyperspectral and deep learning framework was developed to estimate tea quality constituents across seasons and physical states. Samples included field fresh leaves, dried tea leaves, and tea powder, were collected in spring, summer, and autumn. Tea polyphenols and catechins were predicted using original reflectance, harmonic features, and wavelet features fused into multi-domain indices. Extreme gradient boosting, Gaussian process regression, and convolutional neural networks (CNN) were systematically compared to construct the quality estimation models. The result showed that three-feature indices consistently outperformed two-feature indices, yielding R2 from 0.48 to 0.71. CNN achieved the best overall performance among the three modeling approaches, with its optimal accuracy obtained for tea powder samples in autumn, yielding R2 values of 0.81 and 0.76 for tea polyphenols and catechins, respectively. This framework provides an accurate, non-destructive tool for tea quality evaluation and traceability, offering technical support for intelligent agriculture and quality control across the tea industry chain.
In recent years, convolutional-neural-network based stereo matching methods have achieved significant gains compared to conventional methods in terms of both speed and accuracy. Current state-of-the-art disparity estimation algorithms require many parameters and large amounts of computational resources and are not suited for applications on edge devices. In this paper, we propose an end-to-end light-weight network (LWNet) for fast stereo matching, which consists of an efficient backbone with multi-scale feature fusion for feature extraction, a 3D U-Net aggregation architecture for disparity computation, and color guidance in a 2D convolutional neural network (CNN) for disparity refinement. We adopt MobileNetV2 as an efficient backbone in feature extraction. The channel attention module is applied to improve the representational capacity of features and multi-resolution information is adaptively incorporated into the cost volume via cross-scale connections. In addition, instead of using regular 3D convolutions, we utilize pseudo 3D convolutions in the 3D U-Net architecture to aggregate the cost volume for a better balance between computational cost and accuracy. Further, we introduce a left-right consistency check and color guidance and design a robust disparity refinement network with skip connections and dilated convolutions to capture global context information and further improve disparity-estimation accuracy with little computational cost and memory space. A depth-wise separable convolution is proposed to replace all the standard convolutions in the section of disparity refinement, which can decrease computational complexity and the number of parameters without significant accuracy reduction. Extensive experiments on Scene Flow, KITTI 2015, and KITTI 2012 benchmarks demonstrate that the proposed LWNet achieves competitive accuracy when compared with state-of-the-art stereo matching methods.