Longan (Dimocarpus longan Lour.) is highly sensitive to low temperature, which severely restricts its cultivation and industrial development. MYB transcription factors serve as key regulators in plant responses to cold stress. In this study, an R2R3-MYB gene DlMYB108 was cloned from ‘Shixia’ longan. Sequence analysis showed that DlMYB108 contains two typical MYB repeats and shares high homology with cold-responsive MYB108 proteins from other plants. Expression pattern analysis revealed that DlMYB108 is highly expressed in young leaves, which are more sensitive to cold stress, and is significantly induced by low-temperature treatment. Subcellular localization and transcriptional activation assays confirmed that DlMYB108 is a nuclear-localized transcriptional activator. Yeast one-hybrid and dual-luciferase assays demonstrated that DlMYB108 specifically binds to the promoters of DlCBF2 and DlCBF3 and activates their transcription. Heterologous expression of DlMYB108 in Arabidopsis significantly enhanced cold tolerance, accompanied by reduced ion leakage, malondialdehyde (MDA) content and reactive oxygen species (ROS) accumulation, as well as upregulated expression of CBF and cold-responsive genes. Collectively, DlMYB108 positively regulates longan cold tolerance through activating DlCBF2 and DlCBF3 expression, providing a valuable candidate gene for cold-tolerant longan breeding.

Fire detection using deep learning has become a widely adopted approach. However, YOLO-based models often face performance limitations due to NMS, while DETR-based models struggle to meet real-time processing requirements. To address these challenges, we propose FasterGDSF-DETR, a novel fire detection model built upon the RT-DETR framework, designed to enhance both detection accuracy and efficiency. Firstly, this model introduces the FasterDBBNet backbone, which efficiently captures and retains feature information, accelerating the model’s convergence speed. Secondly, we propose the AIFI-GDSF hybrid encoder to reduce information loss in intra-scale interactions and improve the capability of detecting varying morphological flames. Furthermore, to better adapt to complex fire scenarios, we expand the dataset based on the KMU Fire and Smoke database and incorporate WIoU as the loss function to improve model robustness. Experimental results demonstrate that our proposed model surpasses mainstream object detection models in both accuracy and computational efficiency. FasterGDSF-DETR achieves a mean Average Precision of 71.5% on the self-constructed dataset, outperforming the YOLOv9 model of the same scale by 2.4 percentage points. This study introduces a novel task-specific enhancement to the RT-DETR framework, offering valuable insights for future advancements in fire detection technology.

High-speed, precise, and efficient detection of fabric defects is of great significance in enhancing textile industry productivity. Recent advances in Transformer-based deep learning techniques have sparked great interest in their applications in fabric defect detection. Nonetheless, prevailing Transformer-based detection methods struggle to swiftly capture the intricate textures inherent in fabrics, thereby constraining network learning capability. Moreover, as model complexity escalates, these methods are prone to severe overfitting issues. In this research, we introduce a new framework, named ARDSwin-Unet, to tackle these challenges. Firstly, a deformable convolution structure based on shifted window selfattention is designed to extract regions containing rich texture and shape feature information. Secondly, to mitigate overfitting, we propose a Bilevel spatial attention module (BSAM) bottleneck structure and a Res-Drop Swin Transformer block that integrates the residual structure and Dropkey regularization strategy. Finally, experimental results on three different fabric defect datasets demonstrate the superior performance of our proposed model over six state-of-the-art baseline methods across three evaluation metrics.

Crowd counting is a critical aspect of computer vision. This study presents a novel approach to crowd counting by leveraging content-aware upsampling. Existing methods for generating density maps often struggle with accurately capturing local details. To address this issue, we introduce a content-aware upsampling mechanism that assigns a unique upsampling kernel to each feature point. This mechanism enables the model to generate density maps that better align with the distribution characteristics of real density maps. Unlike conventional methods such as bilinear interpolation and transposed convolution, our approach takes into consideration deep semantic information, contextual data, and spatial features, resulting in high-quality density maps. To evaluate the effectiveness of our method, we conducted experiments on three public datasets: NWPU, SHHB, and UCF_QNRF. The results demonstrate the significant improvement in the quality of generated density maps achieved by our algorithm. Moreover, we address labeling errors by using momentum distillation, enhancing the model's ability to process data and reducing the negative impact of labeling noise. In conclusion, our crowd counting method based on content-aware upsampling offers a significant improvement in density map quality. Additionally, the incorporation of momentum distillation helps mitigate labeling errors, enhancing the model's overall performance.

With the advent of various advanced image processing techniques and the increasingly low manipulation threshold, Image Manipulation Detection and Localization (IMDL) has become a complex and challenging task. Previous studies have shown the importance of manipulation artifacts in the IMDL problem. However, existing methods often struggle to capture effective artifacts when addressing with refined forgery (post-processed forgery). In this paper, we introduce MMFA-Net, a novel approach for IMDL. We extract RGB features, Laplacian high-frequency features, Bayar statistical anomaly features, and manipulation imprint from the image through the Multi-modal Feature Extractor module, concatenating them as multi-modal feature inputs to capture manipulation edge artifacts, which are increasingly difficult to detect in image manipulation. In addition, we propose an innovative Multi-head Fuse Attention (MFA) module. Building on the backbone network's ability to model inconsistencies between original and tampered regions at coarse-to-fine granularities, the MFA module can more effectively capture contextual dependencies in the spatial dimensions and inter-channel correlations for more precise localization. The performance on several public datasets demonstrates that the proposed network outperforms existing SOTA models and remains robust against refined forgery.

Lychee is an exotic tropical fruit with a distinct flavor. The genome of cultivar ‘Feizixiao’ was assembled into 15 pseudochromosomes, totaling ~470 Mb. High heterozygosity (2.27%) resulted in two complete haplotypic assemblies. A total of 13,517 allelic genes (42.4%) were differentially expressed in diverse tissues. Analyses of 72 resequenced lychee accessions revealed two independent domestication events. The extremely early maturing cultivars preferentially aligned to one haplotype were domesticated from a wild population in Yunnan, whereas the late-maturing cultivars that mapped mostly to the second haplotype were domesticated independently from a wild population in Hainan. Early maturing cultivars were probably developed in Guangdong via hybridization between extremely early maturing cultivar and late-maturing cultivar individuals. Variable deletions of a 3.7 kb region encompassed by a pair of CONSTANS -like genes probably regulate fruit maturation differences among lychee cultivars. These genomic resources provide insights into the natural history of lychee domestication and will accelerate the improvement of lychee and related crops.

Longan (Dimocarpus longan Lour.) is of great economic significance in South China for its unique taste and nutritional properties. However, longan breeding is mainly based on seedling selection, which generally results in small fruits, low flesh recovery, and few seedless germplasm. Triploid breeding is a central way to improve these problems. In this study, microspore chromosomes were doubled by colchicine and high-temperature treatment to create triploids in longans. The relationship between the development process of male gametophyte of longans and the morphological changes of male flower buds was established. Cytological observation showed that when the male flower buds were in stage I (when the diameter of the flower bud is 1.4–2.0 mm), most of the microspores were at the pachytene to diakinesis stage of meiosis, and the chromosome doubling induction effect was the best at this stage. The results showed that the 2n pollen rate induced by a high temperature of about 38 °C was higher than that induced by colchicine treatment. The highest 2n pollen rate was 5.7% and 5.5% based on the microscopic measurement method and the abnormal separation in tetrad stage estimation method, respectively. Four triploids were successfully obtained from artificial pollination with 2n pollen, with a triploid induction rate of 0.6%. This study will promote ploidy breeding in longan.

"脆丰龙眼"是以"大乌圆龙眼"为母本,"石硖龙眼"为父本的F1杂交后代,经过多年连续观察筛选出的早结丰产大果型龙眼新品种,2021年8月通过了广东省农作物品种评定委员会评定,评定编号为粤评果20210003."脆丰龙眼"果实较大,扁圆形,果皮黄褐色,果肉浅蜡黄,离核,肉质脆嫩化渣,清甜多汁,略有香气;平均单果重13.8g,平均果皮厚度为0.61mm,裂果率较低;可食率69.6％,可溶性固形物含量19.3％,总糖含量16.5g·100g-1,还原糖含量7.4g·100g-1,可滴定酸含量0.25g·kg-1,维生素C含量55.6mg·100g-1."脆丰龙眼"早结丰产,稳产性好,优质大果,在广州地区7月21—31日成熟,在潮汕地区8月1—20日成熟,适宜在龙眼产区推广种植.

华蜜黄皮是以白糖黄皮为母本、郁南无核黄皮为父本进行杂交,从杂交F1代群体中单株优选而成的黄皮新品种.果实鸡心形,果皮橙黄色;肉质细嫩,风味蜜甜,有香气;平均单果质量7.98 g,平均单果种子数1.1粒,可食率69.6％;可溶性固形物含量(w,后同)18.8％,总糖11.8％,还原糖5.4％,可滴定酸0.1％,维生素C602mg·kg-1.华蜜黄皮树势较旺,丰产稳产性较好,早熟,在广州6月底至7月上旬成熟,成熟度较一致,适宜在广东省黄皮产区推广种植.

华蜜2号黄皮是以白糖黄皮为母本、郁南无核黄皮为父本进行杂交,从杂交F,代群体中单株优选而成的黄皮新品种.果实椭圆形,果皮黄褐色;肉质细嫩化渣,蜜甜微酸,香气浓郁;平均单果质量10.08 g,平均单果种子数1.92粒,可食率66.5％;可溶性固形物含量(w,后同)为17.8％,总糖含量11.1％,还原糖含量7.8％,总酸含量1.78％,维生素C含量516mg·kg-1.华蜜2号黄皮树势较旺,树姿开张,丰产稳产性较好,生产示范区平均产量22 500kg·hm-2.中熟,在广州地区7月上中旬成熟,成熟度较一致,适宜在广东省黄皮产区推广种植.

The pathogenic fungus Neoscytalidium dimidiatum (Nd) is the causal agent of pitaya canker and causes significant yield losses. The mechanism by which Nd invades pitaya stems remains largely unknown. Here, quantitative proteomic analysis was employed to investigate pitaya immune responses against Nd infection. A total of 2766 proteins including 244 differentially expressed proteins (DEPs) were identified during infection. Nearly half of the upregulated proteins were predicted to be located in the chloroplast and mitochondrion, implying that these organelles are most affected by fungal infection. Bioinformatics analysis indicated that the DEPs were associated with photosynthesis, phytohormone activity, reactive oxygen species (ROS) homeostasis, and pathogenic defense responses. Notably, the phytohormones auxin and abscisic acid were accumulated for defense against Nd invasion. qRT-PCR validation showed that the mRNA expression levels of auxin binding protein (ABP), auxin response 4 protein (ARP4), and aspartic protease 2 (ASP2) were not consistent with the protein variation, suggesting that these proteins were regulated post-transcriptionally. Additionally, DEPs associated with ROS metabolism changed markedly, indicating that ROS homeostasis in pitaya is important for defense against Nd invasion. In summary, the results revealed the involvement of many essential proteins in the response to Nd infection and provide a basis for studying other biotic stresses of pitaya.

Longan (Dimocarpus longan) is a typical southern subtropical fruit tree species that is sensitive to cold stress. C-repeat binding factors (CBFs), as transcription factors, are crucial components involved in the molecular regulation of the plant response to cold stress. However, the role of CBF homologs in the cold response regulation of longan remains largely unknown. Here, three novel CBF genes, DlCBF1, DlCBF2, and DlCBF3, were cloned from longan. DlCBF1 and DlCBF2 contain an AP2 domain and PKKPAGR and DSAWR CBF signature motifs, while DlCBF3 has mutations within these conserved signature motifs. DlCBF1/2/3 were mainly localized in the nucleus and specifically bound to CRT/DRE cis-elements, resulting in strong transcriptional activation. DlCBF1/2 exhibited tissue expression specificity, and their expression was induced by low temperature, while DlCBF3 had no tissue specificity and barely responded to low temperature. DlCBF1, DlCBF2, and DlCBF3 overexpression in Arabidopsis-enhanced cold tolerance by increasing proline accumulation and reducing reactive oxygen species (ROS) content, accompanied by upregulated expression of cold-responsive genes (AtRD29A, AtCOR15A, AtCOR47, and AtKIN1) in the CBF cold stress response signaling pathway. In conclusion, the biological functions of DlCBF1/2/3 were somewhat conserved, but slow expression of DlCBF1/2 and low expression of DlCBF3 may partly cause the cold sensitivity of longan. Collectively, these results indicated that differences exist in the expression and function of CBF orthologs in the cold-sensitive plant species longan, and these findings may help to improve the understanding of the cold response regulation mechanism and provide important theoretical support for cold-tolerant breeding of longan.

一、“十四五”种业发展面临的机遇与挑战 (一)物种产业形势分析 荔枝龙眼是原产于我国南方的名贵南亚热带水果,无论是品种资源还是栽培面积与产量均居世界之首.我国拥有丰富的荔枝龙眼种质资源,但随着全球经济一体化进程的不断加快和生物技术的迅猛发展,国际间的种业竞争也越来越激烈,呈现出从种质资源主权保护到基因资源产权保护的发展态势.因此,加快推进荔枝龙眼种业发展与科技创新,培育和推广具有自主知识产权的优良品种,加强新品种权保护的立法与监察管理,已成为继续保持我国荔枝龙眼种业国际领先地位、促进我国荔枝龙眼产业可持续发展的迫切需要.

DEGs in â SJâ during floral induction. (XLS 7620 kb)

[目的]流式细胞术是目前测定植物倍性和基因组大小差异的最快、最有效的方法.建立适合荔枝的流式细胞术的方法,对荔枝倍性育种以及基因组大小的确定是必不可少的.[方法]笔者以荔枝的幼嫩叶片为材料,筛选适合荔枝的细胞核提取液配方,建立利用流式细胞仪测定荔枝倍性和基因组大小的方法.用改进的标准两步法,比较了6种常用细胞核提取液提取细胞核的效果.[结果]利用WPB (Woody Plant Buffer)提取的大部分荔枝品种(品系)叶片细胞核稳定、分辨率高、细胞碎片少且细胞G0/G1峰的变异系数(CV)较低,平均CV值为5.12％,说明该配方适用于酚类物质丰富的荔枝细胞核的提取.同时以已知染色体数量的‘无核荔’(2n=30)为外标,检测了18个品种(品系)的倍性,发现参测样品的G0/G1峰与‘无核荔’G0/G1峰的荧光均值的比值为0.78～1.24,说明所测荔枝品种(品系)均为二倍体.以已知基因组大小的‘Stupické polni rané’番茄为内标测定了14个品种(品系)的基因组大小,结果表明荔枝基因大小约为550～620Mb,平均602Mb,不同荔枝品种(品系)间基因组大小存在一定差异.[结论]WPB细胞核提取液提取的荔枝幼嫩的叶片的细胞核质量好,可用于流式细胞术荔枝倍性和基因组大小的测定,测定的结果显示参测荔枝品种(品系)均为二倍体,无单倍或多倍的情况,不同荔枝品种基因组大小有一定的差异.

GO terms significantly overrepresented in the eight enriched profiles of gene expression versus the reference set in ‘SJ’ and ‘SX’. (XLSX 71 kb)

BackgroundLongan (Dimocarpus longan Lour.) is an important fruit tree in the subtropical regions of Southeast Asia and Australia. Among the factors affecting D. longan fruit yield, the difficulty and instability of blossoming is one of the most challenging issues. Perpetual flowering (PF) is a crucial trait for fruit trees and is directly linked to production potential. Therefore, studying the molecular regulatory mechanism of longan PF traits is crucial for understanding and solving problems related to flowering. In this study, comparative transcriptome analysis was performed using two longan cultivars that display opposite flowering phenotypes during floral induction.ResultsWe obtained 853.72M clean reads comprising 125.08 Gb. After comparing these data with the longan genome, 27,266 known genes and 1913 new genes were detected. Significant differences in gene expression were observed between the two genotypes, with 6150 and 6202 differentially expressed genes (DEGs) for SJ' and SX', respectively. The transcriptional landscape of floral transition at the early stage was very different in these two longan genotypes with respect to key hormones, circadian rhythm, sugar metabolism, and transcription factors. Almost all flowering-related DEGs identified are involved in photoperiod and circadian clock pathways, such as CONSTANS-like (COL), two-component response regulator-like (APRRs), gigantea (GI), and early flowering (EFL). In addition, the leafy (LFY) gene, which is the central floral meristem identity gene, may inhibit PF formation in SJ'.ConclusionThis study provides a platform for understanding the molecular mechanisms responsible for changes between PF and seasonal flowering (SF) longan genotypes and may benefit studies on PF trait mechanisms of evergreen fruit trees.

ICE1 (inducer of CBF expression 1) encodes a typical MYC-like basic helix-loop- helix (bHLH) transcription factor that acts as a pivotal component in the cold signalling pathway. In this study, DlICE1, a novel ICE1-like gene, was isolated from the southern subtropical fruit tree longan (Dimocarpus longan Lour.). DlICE1 encodes a nuclear protein with a highly conserved bHLH domain. DlICE1 expression was slightly upregulated under cold stress. Overexpression of DlICE1 in Arabidopsis conferred enhanced cold tolerance via increased proline content, decreased ion leakage, and reduced malondialdehyde (MDA) and reactive oxygen species (ROS) accumulation. Expression of the ICE1-CBF cold signalling pathway genes, including AtCBF1/2/3 and cold-responsive genes (AtRD29A, AtCOR15A, AtCOR47 and AtKIN1), was also significantly higher in DlICE1-overexpressing lines than in wild-type (WT) plants under cold stress. In conclusion, these findings indicate that DlICE1 is a member of the bHLH gene family and positively regulates cold tolerance in D. longan.

Aiming to breed a new cultivar with late maturation and large fruit size, three F1 intergeneric hybrids of longan and litchi populations, viz. 'Shixia' x 'Ziniangxi', Z'aoshu' x 'Ziniangxi' and 'Zhongqiu' x 'Ziniangxi' were evaluated under this experiment. Using the progenies as materials, 30 fruit quality traits including harvesting date, fruit weight, flesh thickness, seed weight, edible: non-edible ratio, soluble solids content, flavor of fruit, etc. were assessed and evaluated. The results showed that there was a significant difference in fruit quality traits among different hybrid progenies. The individual selections were evaluated by the quality weighed average, which was calculated as the scores of three different traits including fruit weight, the edible/non-edible ratio and the soluble solids content. Four hybrids ('Shizi-2', 'Shizi-9', 'Shizi-19' and 'Zaozi-1') were identified as superior individuals with a quality weighed average of more than 83, showing late maturation, large fruit size with high edible/ non-edible ratio. Hence, these individuals will have great value in production application and breeding.

Ubiquitin-conjugating enzymes (E2s or UBC enzymes) play vital roles in plant development and combat various biotic and abiotic stresses. Longan (Dimocarpus longan Lour.) is an important fruit tree in the subtropical region of Southeast Asia and Australia; however the characteristics of the UBC gene family in longan remain unknown. In this study, 40 D. longan UBC genes (DlUBCs), which were classified into 15 groups, were identified in the longan genome. An RNA-seq based analysis showed that DlUBCs showed distinct expression in nine longan tissues. Genome-wide RNA-seq and qRT-PCR based gene expression analysis revealed that 11 DlUBCs were up- or down-regualted in the cultivar "Sijimi" (SJ), suggesting that these genes may be important for flower induction. Finally, qRT-PCR analysis showed that the mRNA levels of 13 DlUBCs under SA (salicylic acid) treatment, seven under methyl jasmonate (MeJA) treatment, 27 under heat treatment, and 16 under cold treatment were up- or down-regulated, respectively. These results indicated that the DlUBCs may play important roles in responses to abiotic stresses. Taken together, our results provide a comprehensive insight into the organization, phylogeny, and expression patterns of the longan UBC genes, and therefore contribute to the greater understanding of their biological roles in longan.