Transcriptome co-expression network analysis identifies key genes and regulators of ripening kiwifruit ester biosynthesis.

BACKGROUND: Aroma is an important organoleptic quality for fruit and has a large influence on consumer preference. Kiwifruit esters undergo rapid and substantial changes contributing to the flavor during fruit ripening. Part of enzymes and their coding genes have been indicated potential candidates for flavor-related esters synthesis. However, there still exist obvious gaps in the biosynthetic pathways of esters and the mechanisms regulating ester biosynthesis in kiwifruit remain unknown. RESULTS: Using gas chromatography-mass spectrometry (GC-MS), volatile compounds of kiwifruit were quantified in response to ethylene (ETH, 100 µl/l, 24 h, 20 °C) and 1-methylcyclopropene (1-MCP, 1 µl/l, 24 h, 20 °C). The results indicated that esters showed the most substantial changes enhanced by ethylene and were inhibited by 1-MCP. Correlations between RNA-seq results and concentrations of esters, constructed using Weighted Gene Co-Expression Network Analysis (WGCNA) indicated that three structural genes (fatty acid desaturase, AdFAD1; aldehyde dehydrogenase, AdALDH2; alcohol acyltransferase, AdAT17) had similar expression patterns that paralled the changes in total ester content, and AdFAD1 transcripts exhibited the highest correlation. In order to search for potential regulators for ester biosynthesis, 14 previously reported ethylene-responsive transcription factors (TFs) were included in the correlation analysis with esters and their biosynthetic genes. Using dual-luciferase assay, the in vivo regulatory activities of TFs on ester biosynthetic gene promoters were investigated and the results indicated that AdNAC5 and AdDof4 (DNA binding with one finger) trans-activated and trans-suppressed the AdFAD1 promoter. CONCLUSIONS: The present study advanced the molecular basis of ripening-related ester biosynthesis in kiwifruit by identifying three biosynthetic related genes AdFAD1, AdALDH2 and AdAT17 by transcriptome analysis, and highlighted the function of two TFs by transactivation studies.

ABA-responsive ABRE-BINDING FACTOR3 activates DAM3 expression to promote bud dormancy in Asian pear.

Bud dormancy is indispensable for the survival of perennial plants in cold winters. Abscisic acid (ABA) has essential functions influencing the endo-dormancy status. Dormancy-associated MADS-box/SHORT VEGETATIVE PHASE-like genes function downstream of the ABA signalling pathway to regulate bud dormancy. However, the regulation of DAM/SVP expression remains largely uncharacterized. In this study, we confirmed that endo-dormancy maintenance and PpyDAM3 expression are controlled by the ABA content in pear (Pyrus pyrifolia) buds. The expression of pear ABRE-BINDING FACTOR3 (PpyABF3) was positively correlated with PpyDAM3 expression. Furthermore, PpyABF3 directly bound to the second ABRE in the PpyDAM3 promoter to activate its expression. Interestingly, both PpyABF3 and PpyDAM3 repressed the cell division and growth of transgenic pear calli. Another ABA-induced ABF protein, PpyABF2, physically interacted with PpyABF3 and disrupted the activation of the PpyDAM3 promoter by PpyABF3, indicating DAM expression was precisely controlled. Additionally, our results suggested that the differences in the PpyDAM3 promoter in two pear cultivars might be responsible for the diversity in the chilling requirements. In summary, our data clarify the finely tuned regulatory mechanism underlying the effect of ABA on DAM gene expression and provide new insights into ABA-related bud dormancy regulation.

Compensating deep focusing distortion for femtosecond laser inscription of low-loss optical waveguides.

Various beam shaping approaches were examined to counter the negative influence of surface aberration arising when inscribing optical waveguides deeply inside of glass with a femtosecond laser. Aberration correction was found unable to completely recover the low-loss waveguide properties, prompting a comprehensive examination of waveguides formed with focused Gaussian-Bessel beams. Diverging conical phase fronts are presented as a hybrid means of partial aberration correction to improve insertion loss and a new, to the best of our knowledge, means of asymmetric beam shaping. In this way, low-loss waveguides are presented over shallow to deep writing depth (2.8 mm) where morphological and modal properties could be further tuned with conical phase front.

PpCBFs selectively regulate PpDAMs and contribute to the pear bud endodormancy process.

KEY MESSAGE: PpCBF2 directly binds to the promoters of PpCBF3 and PpCBF4 to activate their expressions and selectively regulates PpDAMs during the leaf bud endodormancy process of 'Wonhwang' pear (Pyrus pyrifolia). Endodormancy is critical for temperate plant survival under freezing winter conditions, and low temperature is a vital environmental factor in endodormancy regulation. A C-repeat binding factor (CBF) has been found to regulate important DAM transcription factors during endodormancy in pear (Pyrus pyrifolia). In this study, we analyzed the regulation of pear DAM genes by CBFs in further detail. Four CBF and three DAM genes were identified in the pear cultivar 'Wonhwang'. Under natural conditions, PpDAM1 expression decreased from the start of chilling accumulation, while the other two DAM and three CBF genes peaked during endodormancy release. Under chilling treatment, the expressions of PpDAM1, PpDAM2 and PpCBF1 genes were similar to those under natural conditions. Different biochemical methods revealed that PpCBF2/4 can bind to the promoter of PpDAM1 and activate its expression and that PpCBF1/4 can activate PpDAM3. Interestingly, we found that PpCBF2 can activate PpCBF3/4 transcription by directly binding to their promoters. The ICE-CBF regulon is conserved in some plants; three ICE genes were identified in pear, but their expressions did not obviously change under natural and artificial chilling conditions. On the contrary, the selective transcriptional induction of PpCBFs by PpICE1s was observed in a dual-luciferase assay. Considering all these results, we propose that the PpCBF1-PpDAM2 regulon mainly responds to low temperature during endodormancy regulation, with further post-translational regulation by PpICE3. Our results provide basic information on CBF genes functional redundancy and differentiation and demonstrate that the CBF-DAM signaling pathway is involved in the pear bud endodormancy process.

Ultrafast laser burst-train filamentation for non-contact scribing of optical glasses.

A systematic study of glass scribing is presented on the benefits of ultrafast laser burst trains in generating filamentation tracks to guide cleaving of glass substrates. The interplay of Kerr self-focusing, plasma defocusing, and burst-train accumulation effects in filament formation was characterized by time-resolved in-situ microscopic imaging. Various filament-track scribing geometries were compared with and without assistance from burst-train pulse delivery or surface V-groove ablation. The cleaving guidance and reproducibility were examined together with the breaking force, facet morphology and flexural strength of cleaved substrates to assess the overall scribing and cleaving quality. The reported results attest to the benefits and flexibility of burst-mode ultrafast laser interactions to assist cleaving of optically transparent materials along well formed filament arrays.

Phylogenetic, Molecular, and Functional Characterization of PpyCBF Proteins in Asian Pears (Pyrus pyrifolia).

C-repeat binding factor/dehydration-responsive element (CBF/DRE) transcription factors (TFs) participate in a variety of adaptive mechanisms, and are involved in molecular signaling and abiotic stress tolerance in plants. In pear (Pyrus pyrifolia) and other rosaceous crops, the independent evolution of CBF subfamily members requires investigation to understand the possible divergent functions of these proteins. In this study, phylogenetic analysis divided six PpyCBFs from the Asian pear genome into three clades/subtypes, and collinearity and phylogenetic analyses suggested that PpyCBF3 was the mother CBF. All PpyCBFs were found to be highly expressed in response to low temperature, salt, drought, and abscisic acid (ABA) as well as bud endodormancy, similar to PpyCORs (PpyCOR47, PpyCOR15A, PpyRD29A, and PpyKIN). Transcript levels of clade II PpyCBFs during low temperature and ABA treatments were higher than those of clades I and III. Ectopic expression of PpyCBF2 and PpyCBF3 in Arabidopsis enhanced its tolerance against abiotic stresses, especially to low temperature in the first case and salt and drought stresses in the latter, and resulted in lower reactive oxygen species (ROS) and antioxidant gene activities compared with the wild type. The increased expression of endogenous ABA-dependent and -independent genes during normal conditions in PpyCBF2- and PpyCBF3-overexpressing Arabidopsis lines suggested that PpyCBFs were involved in both ABA-dependent and -independent pathways. All PpyCBFs, especially the mother CBF, had high transactivation activities with 6XCCGAC binding elements. Luciferase and Y1H assays revealed the existence of phylogenetically and promoter-dependent conserved CBF-COR cascades in the pear. The presence of a previously identified CCGA binding site, combined with the results of mutagenesis of the CGACA binding site of the PpyCOR15A promoter, indicated that CGA was a core binding element of PpyCBFs. In conclusion, PpyCBF TFs might operate redundantly via both ABA-dependent and -independent pathways, and are strongly linked to abiotic stress signaling and responses in the Asian pear.

Genome wide identification and predicted functional analyses of NAC transcription factors in Asian pears.

BACKGROUND: NAC proteins contribute to diverse plant developmental processes as well as tolerances to biotic and abiotic stresses. The pear genome had been decoded and provided the basis for the genome-wide analysis to find the evolution, duplication, gene structures and predicted functions of PpNAC transcription factors. RESULTS: A total of 185 PpNAC genes were found in pear, of which 148 were mapped on chromosomes while 37 were on unanchored scaffolds. Phylogeny split the NAC genes into 6 clades (Group1- Group6) with their sub clades (~ subgroup A to subgroup H) and each group displayed common motifs with no/minor change. The numbers of exons in each group varied from 1 to 12 with an average of 3 while 44 pairs from all groups showed their duplication events. qPCR and RNA-Seq data analyses in different pear cultivars/species revealed some predicted functions of PpNAC genes i.e. PpNACs 37, 61, 70 (2A), 53, 151(2D), 10, 92, 130 and 154 (3D) were potentially involved in bud endodormancy, PpNACs 61, 70 (2A), 172, 176 and 23 (4E) were associated with fruit pigmentations in blue light, PpNACs 127 (1E), 46 (1G) and 56 (5A) might be related to early, middle and late fruit developments respectively. Besides, all genes from subgroups 2D and 3D were found to be related with abiotic stress (cold, salt and drought) tolerances by targeting the stress responsive genes in pear. CONCLUSIONS: The present genome-wide analysis provided valuable information for understanding the classification, motif and gene structure, evolution and predicted functions of NAC gene family in pear as well as in higher plants. NAC TFs play diverse and multifunctional roles in biotic and abiotic stresses, growth and development and fruit ripening and pigmentation through multiple pathways in pear.

Femtosecond laser filaments for rapid and flexible writing of fiber Bragg grating.

A new beam delivery method is introduced for controlling filament formation in optical fiber that enables point-by-point writing of 1st order fiber Bragg gratings (FBGs) with single femtosecond laser pulses. Uniform filament tracks with azimuthal symmetry were formed fully through the 9.3 µm core waveguide by a modified immersion focusing method to eliminate astigmatism by the cylindrical fiber shape. Filament arrays were precisely assembled inside of single-mode fiber, generating strong FBG resonances in the telecommunication band. Laser exposure control within this unique thin-grating geometry were key to manipulating the relative strength of the Bragg and cladding mode resonances while also independently tailoring their spectral resolution and features. This filament-by-filament writing rapidly forms gratings with highly flexible pattern control to tune wavelength, or introduce optical defects, demonstrated by a π-shifted FBG having a sharp 25 pm resonance embedded within a broader Bragg peak.

Abscisic Acid (ABA ) Promotes the Induction and Maintenance of Pear (Pyrus pyrifolia White Pear Group) Flower Bud Endodormancy.

Dormancy is an adaptive mechanism that allows temperate deciduous plants to survive unfavorable winter conditions. In the present work, we investigated the possible function of abscisic acid (ABA) on the endodormancy process in pear. The ABA content increased during pear flower bud endodormancy establishment and decreased towards endodormancy release. In total, 39 putative genes related to ABA metabolism and signal transductions were identified from pear genome. During the para- to endodormancy transition, PpNCED-2 and PpNCED-3 had high expression levels, while PpCYP707As expression levels were low. However, during endodormancy, the expression of PpCYP707A-3 sharply increased with increasing cold accumulation. At the same time, the ABA content of pear buds declined, and the percentage of bud breaks rapidly increased. On the other hand, the expression levels of PpPYLs, PpPP2Cs, PpSnRK2s, and PpABI4/ABI5s were also changed during the pear flower bud dormancy cycle. Furthermore, exogenous ABA application to para-dormant buds significantly reduced the bud breaks and accelerated the transition to endodormancy. During the whole treatment time, the expression level of PpPP2C-12 decreased to a greater extent in ABA-treated buds than in control. However, the expression levels of PpSnRK2-1, PpSnRK2-4, and PpABI5-1 were higher in ABA-treated buds. Our results indicated that PpCYP707A-3 and PpNCEDs play pivotal roles on the regulation of endodormancy release, while ABA signal transduction pathway also appears to be involved in the process. The present work provided the basic information about the function of ABA-related genes during pear flower bud dormancy process.

Dormancy-associated MADS-box genes and microRNAs jointly control dormancy transition in pear (Pyrus pyrifolia white pear group) flower bud.

Bud dormancy in perennial plants is indispensable to survival over winter and to regrowth and development in the following year. However, the molecular pathways of endo-dormancy induction, maintenance, and release are still unclear, especially in fruit crops. To identify genes with roles in regulating endo-dormancy, 30 MIKC(C)-type MADS-box genes were identified in the pear genome and characterized. The 30 genes were analysed to determine their phylogenetic relationships with homologous genes, genome locations, gene structure, tissue-specific transcript profiles, and transcriptional patterns during flower bud dormancy in 'Suli' pear (Pyrus pyrifolia white pear group). The roles in regulating bud dormancy varied among the MIKC gene family members. Yeast one-hybrid and transient assays showed that PpCBF enhanced PpDAM1 and PpDAM3 transcriptional activity during the induction of dormancy, probably by binding to the C-repeat/DRE binding site, while DAM proteins inhibited the transcriptional activity of PpFT2 during dormancy release. In the small RNA-seq analysis, 185 conserved, 24 less-conserved, and 32 pear-specific miRNAs with distinct expression patterns during bud dormancy were identified. Joint analyses of miRNAs and MIKC genes together with degradome data showed that miR6390 targeted PpDAM transcripts and degraded them to release PpFT2. Our data show that cross-talk among PpCBF, PpDAM, PpFT2, and miR6390 played important roles in regulating endo-dormancy. A model for the molecular mechanism of dormancy transition is proposed: short-term chilling in autumn activates the accumulation of CBF, which directly promotes DAM expression; DAM subsequently inhibits FT expression to induce endo-dormancy, and miR6390 degrades DAM genes to release endo-dormancy.

Burst train generator of high energy femtosecond laser pulses for driving heat accumulation effect during micromachining.

A new method for generating high-repetition-rate (12.7-38.2 MHz) burst trains of femtosecond laser pulses has been demonstrated for the purpose of tailoring ultrashort laser interactions in material processing that can harness the heat accumulation effect among pulses separated by a short interval (i.e., 26 ns). Computer-controlled time delays were applied to synchronously trigger the high frequency switching of a high voltage Pockels cell to specify distinctive values of polarization rotation for each round-trip of a laser pulse cycling within a passive resonator. Polarization dependent output coupling facilitated the flexible shaping of the burst envelope profile to provide burst trains of up to ∼1 mJ of burst energy divided over a selectable number (1 to 25) of pulses. Individual pulses of variable energy up to 150 µJ and with pulse duration tunable over 70 fs to 2 ps, were applied in burst trains to generate deep and high aspect ratio holes that could not form with low-repetition-rate laser pulses.

Few-Shot Learning With Enhancements to Data Augmentation and Feature Extraction.

The few-shot image classification task is to enable a model to identify novel classes by using only a few labeled samples as references. In general, the more knowledge a model has, the more robust it is when facing novel situations. Although directly introducing large amounts of new training data to acquire more knowledge is an attractive solution, it violates the purpose of few-shot learning with respect to reducing dependence on big data. Another viable option is to enable the model to accumulate knowledge more effectively from existing data, i.e., improve the utilization of existing data. In this article, we propose a new data augmentation method called self-mixup (SM) to assemble different augmented instances of the same image, which facilitates the model to more effectively accumulate knowledge from limited training data. In addition to the utilization of data, few-shot learning faces another challenge related to feature extraction. Specifically, existing metric-based few-shot classification methods rely on comparing the extracted features of the novel classes, but the widely adopted downsampling structures in various networks can lead to feature degradation due to the violation of the sampling theorem, and the degraded features are not conducive to robust classification. To alleviate this problem, we propose a calibration-adaptive downsampling (CADS) that calibrates and utilizes the characteristics of different features, which can facilitate robust feature extraction and benefit classification. By improving data utilization and feature extraction, our method shows superior performance on four widely adopted few-shot classification datasets.

Self-Supervised Monocular Depth Estimation With Self-Perceptual Anomaly Handling.

It is attractive to extract plausible 3-D information from a single 2-D image, and self-supervised learning has shown impressive potential in this field. However, when only monocular videos are available as training data, moving objects at similar speeds to the camera can disturb the reprojection process during training. Existing methods filter out some moving pixels by comparing pixelwise photometric error, but the illumination inconsistency between frames leads to incomplete filtering. In addition, existing methods calculate photometric error within local windows, which leads to the fact that even if an anomalous pixel is masked out, it can still implicitly disturb the reprojection process, as long as it is in the local neighborhood of a nonanomalous pixel. Moreover, the ill-posed nature of monocular depth estimation makes the same scene correspond to multiple plausible depth maps, which damages the robustness of the model. In order to alleviate the above problems, we propose: 1) a self-reprojection mask to further filter out moving objects while avoiding illumination inconsistency; 2) a self-statistical mask method to prevent the filtered anomalous pixels from implicitly disturbing the reprojection; and 3) a self-distillation augmentation consistency loss to reduce the impact of ill-posed nature of monocular depth estimation. Our method shows superior performance on the KITTI dataset, especially when evaluating only the depth of potential moving objects.

Exogenous BR delayed peach fruit softening by inhibiting pectin degradation enzyme genes.

Peach fruit deteriorates and senesces rapidly when stored at room temperature. Brassinosteroids (BRs) play an important role in regulating plant growth and development and maintaining fruit quality. However, little information is available on the effect of BRs on the senescence of harvested peach fruit. In this study, different concentrations of BR were used to treat 'Hongniang' peach fruit, and the results showed that 10 µM BR was the most beneficial concentration to delay the senescence of peach fruits. BR treatment delayed the decrease of fruit firmness, the release of ethylene, the increase in water-soluble pectin (WSP) and ionic-soluble pectin (ISP) content and the decrease in covalently bound pectin (CBP) content, inhibited the activities of pectin degradation enzymes, and inhibited the gene expression of PpPME1/3, PpPG, PpARF2, and PpGAL2/16. In addition, BR treatment also inhibited the expression of PpBES1-5/6. Cis-acting regulatory element analysis of pectin degradation enzyme promoters showed that many of them contained BES1 binding elements. All the above results showed that BR treatment had a positive effect on delaying the senescence of peach fruit and prolonging its storage period.

The Malus domestica metal tolerance protein MdMTP11.1 was involved in the detoxification of excess manganese in Arabidopsis thaliana.

Ion homeostasis is maintained in plant cells by specialized transporters. However, functional studies on Mn transporters in apple trees have not been reported. MdMTP11.1, which encodes a putative Mn-MTP transporter in Malus domestica, was expressed highly in leaves and induced by Mn stress. Subcellular localization analysis of the MdMTP11.1-GFP fusion protein indicated that MdMTP11.1 was targeted to the Golgi. Meanwhile, overexpression of MdMTP11.1 in Arabidopsis thaliana conferred increased resistance to plants under toxic Mn levels, as evidenced by increased biomass of whole plant and length of primary root. Analysis of Mn bioaccumulation indicated that overexpression of MdMTP11.1 effectively reduced the content of Mn in every subcellular component and chemical forms when the plants were subjected with Mn stress. The majority of Mn of action were bound to cell wall and combined with un-dissolved phosphate. Besides, contents of malondialdehyde (MDA), proline and hydrogen peroxide (H2O2) were significantly lower, while content of chlorophyll and activities of CAT, SOD, POD and APX were significantly higher in MdMTP11.1-over-expressing plants compared with that in wild type plants under Mn stress. Taken together, these results suggest that MdMTP11.1 is a Mn specific transporter localized to the Golgi can maintain the phenotype, reduce the Mn accumulation and alleviate damage of oxidative stress, conferring the positive role of Mn tolerance.

Low temperatures inhibit the pectin degradation of 'Docteur Jules Guyot' pear (Pyrus communis L.).

The most remarkable characteristic of European pears is extremely perishable and difficult to store after postharvest softening. Low-temperature storage is one of the most commonly used methods to prolong the shelf life of European pears. However, the regulatory mechanism of the low-temperature delay of the softening of European pears is still unclear. In this study, the fruit firmness, pectin polysaccharide content, pectin-degrading enzyme activity, and pectin degradation gene expression of 'Docteur Jules Guyot' pears under low temperature (LT) and room temperature (RT) were analyzed. It was found that water-soluble pectin (WSP) was significantly negatively correlated with fruit flesh firmness, and the activities of several pectin-degrading enzymes were inhibited under LT storage conditions. In addition, it was also found that the gene expression patterns of PcPME2, PcPME3, PcPG1, PcPG2, PcPL, PcGAL1, PcGAL2, PcGAL4, and PcARF1 were inhibited by LT. The C-repeat binding factors PcCBF1 and PcCBF2 were also inhibited by long-term LT storage. Correlation analysis showed that the expression of PcCBFs was positively correlated with pectin-degradation enzyme genes, and we found that the promoters of many pectin-degradation enzyme genes contain the CRT/DRE motif, which CBF can directly bind. Therefore, it is speculated that long-term low-temperature conditions inhibit pectin degradation through PcCBFs.

Low temperature inhibits pectin degradation by PpCBFs to prolong peach storage time.

Low-temperature storage is a widely used method for peach fruit storage. However, the impact of PpCBFs on pectin degradation during low-temperature storage is unclear. As such, in this study, we stored the melting-flesh peach cultivar "Fuli" at low temperature (LT, 6°C) and room temperature (RT, 25°C) to determine the effect of different temperatures on its physiological and biochemical changes. Low-temperature storage can inhibit the softening of "Fuli" peaches by maintaining the stability of the cell wall. It was found that the contents of water-soluble pectin and ionic-soluble pectin in peach fruit stored at RT were higher than those stored at LT. The enzyme activities of polygalacturonase (PG), pectate lyase (PL), and pectin methylesterase (PME) were all inhibited by LT. The expressions of PpPME3, PpPL2, and PpPG were closely related to fruit firmness, but PpCBF2 and PpCBF3 showed higher expression levels at LT than RT. The promoters of PpPL2 and PpPG contain the DER motif, which suggested that PpCBF2 and PpCBF3 might negatively regulate their expression by directly binding to their promoters. These results indicated that LT may maintain firmness by activating PpCBFs to repress pectin-degradation-related enzyme genes during storage.

Genome-wide characterization of SINA E3 ubiquitin ligase family members and their expression profiles in response to various abiotic stresses and hormones in kiwifruit.

SINA (Seven in absentia) proteins in the subtype of E3 ubiquitin ligase family have important functions in regulating the growth and development as well as in response to abiotic and biotic stresses in plants. However, the characteristics and possible functions of SINA family proteins in kiwifruit are not studied. In this research, a total number of 11 AcSINA genes in the kiwifruit genome were identified. Chromosome location and multiple sequence alignment analyses indicated that they were unevenly distributed on 10 chromosomes and all contained the typical N-terminal RING domain and C-terminal SINA domain. Phylogenetic, gene structure and collinear relationship analyses revealed that they were highly conserved with the same gene structure, and have gone through segmental duplication events. Expression pattern analyses demonstrated that all AcSINAs were ubiquitously expressed in roots, stems and leaves, and were responsive to different abiotic and plant hormone treatments with overlapped but distinct expression patterns. Further yeast two-hybrid and Arabidopsis transformation analyses demonstrated most AcSINAs interacted with itself or other AcSINA members to form homo- or heterodimers, and ectopic expression of AcSINA2 in Arabidopsis led to hypersensitive growth phenotype of transgenic seedlings to ABA treatment. Our results reveal that AcSINAs take part in the response to various abiotic stresses and hormones, and provide important information for the functional elucidation of AcSINAs in vine fruit plants.

Heterologous expression of ISU1 gene from Fragaria vesca enhances plant tolerance to Fe depletion in Arabidopsis.

Iron-sulfur (Fe-S) cluster assembly genes play important roles in plant growth and development. However, their biological function in fruit crops is still unknown, especially in strawberry. In this study, Fe depletion significantly inhibited the growth, photosynthesis, Fe accumulation level and the enzyme activity of Fe-S proteins of aconitase (ACO), nitrate reductase (NiR) and succinate dehydrogenase (SDH) in strawberry seedlings. In addition, 40 Fe-S cluster assembly genes were isolated from strawberry, which were significantly varied among different tissues/organs and were differentially responded to Fe depletion in different tissue parts. In total, 79% of the responsive genes were up-regulated in shoots, while 65% of the responsive genes were down-regulated in roots under Fe depletion. Moreover, the expression level of ISU1 was the highest in strawberry tissues, especially in young fruits, and over-expression of ISU1 gene in Arabidopsis significantly enhanced the Fe accumulation, leaf total chlorophyll, ACO and SDH activities in transgenic lines, and strengthened plant tolerance to Fe depletion. This study provides gene resources to elucidate the molecular mechanisms of Fe-S cluster assembly in strawberry, and lays a theoretical foundation to reveal Fe nutrition and metabolism in Rosaceae fruits.

Self-Supervised Monocular Depth Estimation With Multiscale Perception.

Extracting 3D information from a single optical image is very attractive. Recently emerging self-supervised methods can learn depth representations without using ground truth depth maps as training data by transforming the depth prediction task into an image synthesis task. However, existing methods rely on a differentiable bilinear sampler for image synthesis, which results in each pixel in a synthetic image being derived from only four pixels in the source image and causes each pixel in the depth map to perceive only a few pixels in the source image. In addition, when calculating the photometric error between a synthetic image and its corresponding target image, existing methods only consider the photometric error within a small neighborhood of each single pixel and therefore ignore correlations between larger areas, which causes the model to tend to fall into the local optima for small patches. In order to extend the perceptual area of the depth map over the source image, we propose a novel multi-scale method that downsamples the predicted depth map and performs image synthesis at different resolutions, which enables each pixel in the depth map to perceive more pixels in the source image and improves the performance of the model. As for the locality of photometric error, we propose a structural similarity (SSIM) pyramid loss to allow the model to sense the difference between images in multiple areas of different sizes. Experimental results show that our method achieves superior performance on both outdoor and indoor benchmarks.