Plant Methods最新文献

Background: Arabidopsis thaliana is the leading model plant used to study plant-pathogen interactions. High-throughput phenotyping allows for the simultaneous study of many plants with high-frequency image acquisition. Nevertheless, the segmentation of symptomatic plants on natural soil remains challenging, requiring the annotation of hundreds of images and the subsequent training of specialized models for each pathosystem considered. This paper presents a novel approach to segmenting A. thaliana plants' time series using a single annotated image.

Results: Images of A. thaliana plants infected with Pseudomonas syringae pathovar tomato strain DC3000 were annotated with precise segmentation masks. We compared various mask segmentation methods; our one-shot learning approach obtained a Dice score of 0.977 on our test dataset. Variables extracted from the segmented images allowed statistical discrimination between infected and control plants. We used our one-shot learning approach without further fine-tuning on a new pathosystem; A. thaliana infected with Ralstonia pseudosolanacearum, strain GMI1000. We obtained a Dice score of 0.966 in the second test dataset. We also obtained a Pearson correlation coefficient of - 0.928 between the annotated quantitative disease index and the variable generated with our method.

Conclusions: This work provides a pipeline to segment symptomatic A. thaliana plants by leveraging a visual foundation model. The method has been used successfully on two different pathogens, is fast to train, and does not need a large dedicated graphical processing unit. Our method has characterized plant-pathogen interactions of two pathosystems without fine-tuning for the second pathosystem. Its ease of use and low computing requirements should make adapting our approach to other high-throughput phenotyping platforms easy.

In recent years, non-destructive and non-invasive methods for 3D plant reconstruction have gained increasing importance in plant phenotyping. Morphological traits reflect the physiological status of a plant and serve as key indicators for precision agriculture, crop protection, and food quality assessment. Accurate and efficient 3D modelling enables objective and repeatable monitoring of plant development and health, thus supporting data-driven decision-making in agricultural and food research. This study presents a novel, cost-effective, and flexible photogrammetric apparatus for the routine analysis of plant morphological traits under controlled laboratory conditions. Existing systems often rely on expensive instrumentation and provide limited adaptability, whereas the platform described here combines affordability with high precision and robustness. A key innovation is the use of a robotic arm to control an industrial RGB camera, providing substantial flexibility in image acquisition. This mobility ensures comprehensive coverage of plants of different sizes and architectures while minimising occlusions. Another distinctive feature is the implementation of an optimised parameter tweak in the photogrammetric pipeline, which markedly improves the reconstruction of thin and delicate plant parts such as leaves, petioles, and fine stems. In combination with optimised acquisition parameters, including an exposure time of 50 milliseconds, a tweak value of 0.9, and a camera-to-object distance of 16 centimetres, the system achieves consistent model fidelity across diverse plant structures. Efficiency was further enhanced through automation and an optimised scanning procedure. Comparative testing showed that using a larger number of camera positions with fewer frames per position improved throughput, with the best configuration consisting of three height levels and 40 frames each. These improvements reduced the processing time by 75%, decreasing the average scan duration from 8 min to only 2.7 min per plant, while maintaining accuracy and reliability. Overall, the developed apparatus constitutes a reliable and low-cost solution that integrates robotic-assisted flexibility, improved reconstruction through the parameter tweak, and markedly reduced scanning time. The combination of precision, affordability, and efficiency makes the system competitive with existing approaches and, due to its accessibility and detailed methodological description, provides a distinctive contribution to the phenotyping community.

The generation of a large amount of ground truth data is an essential bottleneck for the application of deep learning-based approaches to plant image analysis. In particular, the generation of accurately labeled images of various plant types at different developmental stages from multiple renderings is a laborious task that substantially extends the time required for AI model development and adaptation to new data. Here, generative adversarial networks (GANs) can potentially offer a solution by enabling widely automated synthesis of realistic images of plant and background structures. In this study, we present a two-stage GAN-based approach to generation of pairs of RGB and binary-segmented images of greenhouse-grown plant shoots. In the first stage, FastGAN is applied to augment original RGB images of greenhouse-grown plants using intensity and texture transformations. The augmented data were then employed as additional test sets for a Pix2Pix model trained on a limited set of 2D RGB images and their corresponding binary ground truth segmentation. This two-step approach was evaluated on unseen images of different greenhouse-grown plants. Our experimental results show that the accuracy of GAN predicted binary segmentation ranges between 0.88 and 0.95 in terms of the Dice coefficient. Among several loss functions tested, Sigmoid Loss enables the most efficient model convergence during the training achieving the highest average Dice Coefficient scores of 0.94 and 0.95 for Arabidopsis and maize images. This underscores the advantages of employing tailored loss functions for the optimization of model performance.

Background: The use of Illumina sequencing technologies has enabled the identification and removal of mutations in various plant species. However, the Illumina sequencing method requires a considerable amount of data to ensure its integrity and quality due to the enormous number of false positives. This study aimed to explore an effective genomic data analysis for the detection of heterozygous variant (HV) in rice varieties.

Results: We compared the accuracy of four combinations of mapping tools and variant calling pipelines and selected BWA-MEM2 with GATK4.3 HaplotypeCaller. To detect heterozygous de novo polymorphisms such as HVs in the three different rice varieties (Nipponbare, Kitaake, and Hinohikari), we adopted the following cost-saving procedures; secondary references were created in Nipponbare and Kitaake, and generation-based comparison was performed in Hinohikari. The similar HVs were estimated by the three varieties to range from 2.55814 × 10^-8 to 4.41860 × 10^-8, with an average of 3.10278 × 10^-8 per nucleotide in a single rice plant, a rate consistent with observations in other organisms. Of 107 HVs identified in all eight plant samples, nine were found to be non-synonymous, resulting in an average of one non-synonymous HV per plant in a single generation.

Conclusions: We have developed a methodology for the detection of true positive HVs within Illumina sequencing techniques. This system removed false positive HVs, allowing for the estimation of true positive HVs and, consequently, the estimation of the mutation rate. The study outlines a clear, step-by-step procedure that can be employed to detect true HVs in different organisms.

Background: Agriculture plays a pivotal role in global food security and socio-economic stability, yet crop productivity remains threatened by plant diseases that incur substantial economic losses. Pomegranate is an important fruit for both nutrition and business, but it is easily infected by pathogens that can lower yields by 20 to 40 percent. Traditional methods of finding these pathogens by hand are time-consuming, subjective, and not very effective, while existing deep learning models struggle with field noise, lighting variations, and computational inefficiency. To address these challenges, this study proposes an automated framework integrating a modified ResNet101 architecture with a Hybrid Genetic Algorithm-Particle Swarm Optimization (HGA-PSO) method. The approach employs dual-stream processing of original and noise-augmented images (Gaussian, salt-and-pepper, speckle) to enhance robustness.

Results: The framework achieved exceptional performance on a dataset of 5,000 images across five classes (four diseases, one healthy). Feature fusion from dual streams and HGA-PSO optimization reduced dimensionality by 50-70% while preserving discriminative power. Under rigorous 5-fold cross-validation, the Multi-Layer Perceptron (MLP) classifier attained 99.10% accuracy, a perfect ROC-AUC score (1.00), and high precision-recall metrics. Confusion matrices revealed near-zero misclassification, and real-world tests (single/batch images) confirmed strong generalization. Grad-CAM + + visualizations validated precise localization of disease regions. The model outperformed existing techniques (e.g., PSO-YOLOv8: 98.86%, Transformer models: 93.13%) in accuracy, precision, recall, and F1-score CONCLUSIONS: This research presents an optimized model for pomegranate disease detection by combining deep learning with nature inspired optimization. The dual-stream feature fusion and HGA-PSO significantly improves robustness again environment variability while reducing computation overhead. This framework offers a scalable solution for precision agriculture, enabling early disease intervention to mitigate economic losses. Future research could improve scalability and usefulness by looking into lightweight optimization methods, model interpretability, and how they can be used in limited-resource agricultural settings.

Background: Accurate quantification of RNA isoforms is critical for understanding gene regulation. However, conventional reverse transcription-quantitative real-time PCR (RT-qPCR) has limited sensitivity for low-abundance transcript isoforms, as quantification cycle (Cq) values above 30 are often considered unreliable. While transcriptome-wide analyses can address this limitation, they require costly deep sequencing and complex bioinformatics. Moreover, isoform-specific qPCR is often confounded by differential primer efficiency when comparing similar transcripts.

Results: To overcome the sensitivity and amplification bias limitations of conventional RT-qPCR for detecting known low-abundance and alternatively spliced transcripts, we developed STALARD (Selective Target Amplification for Low-Abundance RNA Detection), a rapid (< 2 h) and targeted two-step RT-PCR method using standard laboratory reagents. STALARD selectively amplifies polyadenylated transcripts sharing a known 5'-end sequence, enabling efficient quantification of low-abundance isoforms. When applied to Arabidopsis thaliana, STALARD successfully amplified the low-abundance VIN3 transcript to reliably quantifiable levels. Amplification of FLM, MAF2, EIN4, and ATX2 isoforms by STALARD reflected known splicing changes during vernalization, including cases where conventional RT-qPCR failed to detect relevant isoforms. STALARD also enabled consistent quantification of the extremely low-abundance antisense transcript COOLAIR, resolving inconsistencies reported in previous studies. In combination with nanopore sequencing, STALARD further revealed novel COOLAIR polyadenylation sites not captured by existing annotations.

Conclusion: STALARD provides a sensitive, simple, and accessible method for isoform-level quantification of low-abundance transcripts that share a known 5'-end sequence. Its compatibility with both qPCR and long-read sequencing makes it a versatile tool for analyzing transcript variants and identifying previously uncharacterized 3'-end structures, provided that isoform-specific 5'-end sequences are known in advance.

Background: Paeonia ostii, an economically important oil-producing peony cultivar, faces challenges in large-scale cultivation due to low propagation rates and long cultivation cycles. This study aimed to optimize tissue culture protocols for P. ostii 'Fengdan No. 3' by evaluating vernalization and etiolation pretreatments on single-node and leaf explants.

Results: Vernalization and etiolation treatments significantly enhanced in vitro regeneration of P. ostii, resulting in improved organogenic responses and reduced browning. Optimal sterilization and culture conditions were established for both single-node and leaf explants. For single-node explants, NN69 medium delivered the highest shoot induction rate (66.7%) with moderate browning. Supplementation with 0.1 mg·L⁻¹ indole-3-butyric acid (IBA) and 0.2 mg·L⁻¹ N-(2-chloro-4-pyridyl)-N'-phenylurea (CPPU) further enhanced shoot multiplication (4.5-fold) without hyperhydricity. The addition of white-red light increased shoot elongation to 2.27 cm. For leaf explants, callus induction reached 67.8% under 0.3 mg·L⁻¹ IBA and 0.9 mg·L⁻¹ CPPU, while shoot induction peaked at 54.4% with 0.2 mg·L⁻¹ IBA and 0.2 mg·L⁻¹ CPPU, without browning. The incorporation of 0.2 mg·L⁻¹ IBA and 3 mg·L⁻¹ CaCl₂ in the rooting medium promoted rapid adventitious root formation (60%) with robust, non-browning roots systems.

Conclusion: This study established an effective tissue culture platform for P. ostii by integrating vernalization-etiolation pretreatment with optimized culture conditions. This platform addresses the limitations of conventional propagation methods and offers a foundation for large-scale clonal propagation and future genetic improvement of this valuable species.

In rice pest management, accurate pest detection is critical for intelligent agricultural systems, yet challenges like limited dataset availability, pest occlusion, and insufficient small object detection accuracy hinder effective monitoring. To address the aforementioned challenges, this study presents YOLO-PEST, an innovative detection approach based on the YOLOv5s architecture to address these issues. YOLO-PEST collects rice pest images from multiple channels and images are randomly cropped to occlude detection boxes, effectively simulating pest overlapping scenarios. During the feature fusion process, the ConvNeXt module is integrated to improve the detection accuracy for small objects via multiscale feature extraction. Additionally, the CoTAttention mechanism is incorporated to enhance the model's robustness under complex environmental conditions. Comparative experiments show that the YOLO-PEST approach achieves a 97% of mAP@0.5, representing a 1.4-point improvement compared with previous methods, thus verifying its effectiveness in rice pest management.

Background: Protein detection on large somatic chromosomes typically includes paraformaldehyde fixation and squashing of enzymatically softened root tips in a buffer. It often suffers from chromosome clumping, poor chromosome morphology, non-specific fluorescence, insufficient immunoreactivity, which collectively reduce the credibility of immunolabeling, hindering its effective combination with fluorescence in situ hybridization (FISH). Material harvesting and pre-detection steps must be completed within a short time, usually one day, which complicates research. The aim of this study was to develop a simple efficient squash-based protocol for technically demanding formaldehyde-fixed large chromosomes/nuclei (Allium, Scilla, Tradescantia), that ensures: long-term storage of the fixed root tips and of slide preparations, the obtaining of high-quality immunolabeled metaphase plates/nuclear spreads with no or minimal unspecific fluorescence and running a sensitive immunoFISH-karyotyping.

Results: Fixation with 10% buffered formalin was combined with prolonged or overnight storage of the fixed intact tissue in 70% ethanol, digestion with pectinase-cellulase mix in citrate buffer, moderate squashing of root tip tissues in 45% acetic acid, slide freezing followed by ethanol-aided cell adherence to a slide, storage of the preparations in glycerin, one-two cycles of microwave antigen retrieval (MWAR). This resulted in optimal chromosomal/nuclear spreading, good cell adherence to the slide, effective antigen retrieval, reduced/eliminated non-specific fluorescence, good penetration of antibodies. The MWAR-assisted protein redetection could have been performed to strengthen the signals. The protocol was compatible with FISH to perform a sensitive immunoFISH with the rDNA probe and simultaneous visualization of FISH-signals and protein foci.

Conclusion: As a novel approach, the protocol includes an array of steps and options not described in chromosomal immunoprotocols that used aldehyde-fixed root tips for squashing, e.g., fixation with neutral-buffered formalin, storage of root tips in ethanol, squash in acetic acid, MWAR, protein redetection, immunoFISH-aided simultaneous DNA-protein visualization. It ensures chromosomal/nuclear spread of exceptional quality, rapid preparation of the fixing solution, prolonged storage of both fixed tissues and slide preparations, epitope redetection, sensitive immunoFISH-karyotyping. The described methodology provides unprecedented flexibility in laboratory work and significantly expands plant cyto-epigenetic research.

Tomato early blight, caused by Alternaria solani, poses a significant threat to crop yields. Existing detection methods often struggle to accurately identify small or multi-scale lesions, particularly in early stages when symptoms exhibit low contrast and only subtle differences from healthy tissue. Blurred lesion boundaries and varying degrees of severity further complicate accurate detection. To address these challenges, we present YOLOv11-AIU, a lightweight object detection model built on an enhanced YOLOv11 framework, specifically designed for severity grading of tomato early blight. The model integrates a C3k2_iAFF attention fusion module to strengthen feature representation, an Adown multi-branch downsampling structure to preserve fine-scale lesion features, and a Unified-IoU loss function to enhance bounding box regression accuracy. A six-level annotated dataset was constructed and expanded to 5,000 images through data augmentation. Experimental results demonstrate that YOLOv11-AIU outperforms models such as YOLOv3-tiny, YOLOv8n, and SSD, achieving a mAP@50 of 94.1%, mAP@50-95 of 93.4%, and an inference speed of 15.67 FPS. When deployed on the Luban Cat5 platform, the model achieved real-time performance, highlighting its strong potential for practical, field-based disease detection in precision agriculture and intelligent plant health monitoring.