Cooperation of local features and global representations by a dual-branch network for transcription factor binding sites prediction.

Interactions between DNA and transcription factors (TFs) play an essential role in understanding transcriptional regulation mechanisms and gene expression. Due to the large accumulation of training data and low expense, deep learning methods have shown huge potential in determining the specificity of TFs-DNA interactions. Convolutional network-based and self-attention network-based methods have been proposed for transcription factor binding sites (TFBSs) prediction. Convolutional operations are efficient to extract local features but easy to ignore global information, while self-attention mechanisms are expert in capturing long-distance dependencies but difficult to pay attention to local feature details. To discover comprehensive features for a given sequence as far as possible, we propose a Dual-branch model combining Self-Attention and Convolution, dubbed as DSAC, which fuses local features and global representations in an interactive way. In terms of features, convolution and self-attention contribute to feature extraction collaboratively, enhancing the representation learning. In terms of structure, a lightweight but efficient architecture of network is designed for the prediction, in particular, the dual-branch structure makes the convolution and the self-attention mechanism can be fully utilized to improve the predictive ability of our model. The experiment results on 165 ChIP-seq datasets show that DSAC obviously outperforms other five deep learning based methods and demonstrate that our model can effectively predict TFBSs based on sequence feature alone. The source code of DSAC is available at https://github.com/YuBinLab-QUST/DSAC/.

DeepSTF: predicting transcription factor binding sites by interpretable deep neural networks combining sequence and shape.

Precise targeting of transcription factor binding sites (TFBSs) is essential to comprehending transcriptional regulatory processes and investigating cellular function. Although several deep learning algorithms have been created to predict TFBSs, the models' intrinsic mechanisms and prediction results are difficult to explain. There is still room for improvement in prediction performance. We present DeepSTF, a unique deep-learning architecture for predicting TFBSs by integrating DNA sequence and shape profiles. We use the improved transformer encoder structure for the first time in the TFBSs prediction approach. DeepSTF extracts DNA higher-order sequence features using stacked convolutional neural networks (CNNs), whereas rich DNA shape profiles are extracted by combining improved transformer encoder structure and bidirectional long short-term memory (Bi-LSTM), and, finally, the derived higher-order sequence features and representative shape profiles are integrated into the channel dimension to achieve accurate TFBSs prediction. Experiments on 165 ENCODE chromatin immunoprecipitation sequencing (ChIP-seq) datasets show that DeepSTF considerably outperforms several state-of-the-art algorithms in predicting TFBSs, and we explain the usefulness of the transformer encoder structure and the combined strategy using sequence features and shape profiles in capturing multiple dependencies and learning essential features. In addition, this paper examines the significance of DNA shape features predicting TFBSs. The source code of DeepSTF is available at https://github.com/YuBinLab-QUST/DeepSTF/.

Fine-tuned local coordination environment of Pt-N in nanocarbons for efficient propane dehydrogenation.

Based on the disturbance of electronic density, nitrogen-doped nanocarbons show promising properties to anchor metal clusters. However, precisely regulating the coordination mode between N species and the active site remains challenging. Herein, we rationally designed three N types (graphitic N, pyridinic N and pyrrolic N) in nanocarbons to anchor Pt clusters for the benchmark propane dehydrogenation. The specific activity of the pyridinic-N-doped catalyst was 147.54 molC3H6 molPt-1 h-1 at 550 °C, which was 1.3 times higher than those of graphitic- and pyrrolic-N-doped catalysts. Unlike the regular tetrahedron Pt cluster in the graphitic-N catalyst or the distorted three-layered Pt cluster in the pyrrolic-N catalyst, the Pt cluster in the pyridinic-N catalyst was an inverted tetrahedron, which increased the contact degree without geometric repulsion towards C-H bond scission. The geometric parameters of detached H and C atoms in the methylene group for the pyridinic N catalyst was decreased to strengthen the C-H bond scission. After CH3CHCH3* adsorption, the Bader charge of the Pt active site also became highly positive, which tailored the d-band center closer to the Fermi level and provided more vacant orbitals for C-H bond breakage. Therefore, pyridinic N in nanocarbons is promising to anchor small-sized Pt for alkane dehydrogenation in terms of geometric and electronic effects.

Research on Wind Turbine Fault Detection Based on the Fusion of ASL-CatBoost and TtRSA.

The internal structure of wind turbines is intricate and precise, although the challenging working conditions often give rise to various operational faults. This study aims to address the limitations of traditional machine learning algorithms in wind turbine fault detection and the imbalance of positive and negative samples in the fault detection dataset. To achieve the real-time detection of wind turbine group faults and to capture wind turbine fault state information, an enhanced ASL-CatBoost algorithm is proposed. Additionally, a crawling animal search algorithm that incorporates the Tent chaotic mapping and t-distribution mutation strategy is introduced to assess the sensitivity of the ASL-CatBoost algorithm toward hyperparameters and the difficulty of manual hyperparameter setting. The effectiveness of the proposed hyperparameter optimization strategy, termed the TtRSA algorithm, is demonstrated through a comparison of traditional intelligent optimization algorithms using 11 benchmark test functions. When applied to the hyperparameter optimization of the ASL-CatBoost algorithm, the TtRSA-ASL-CatBoost algorithm exhibits notable enhancements in accuracy, recall, and other performance measures compared with the ASL-CatBoost algorithm and other ensemble learning algorithms. The experimental results affirm that the proposed algorithm model improvement strategy effectively enhances the wind turbine fault detection classification recognition rate.

Hydrodeoxygenation of Oxygen-Containing Aromatic Plastic Wastes to Liquid Organic Hydrogen Carriers.

To address the global plastic pollution issues and the challenges of hydrogen storage and transportation, we report a system, based on the hydrodeoxygenation (HDO) of oxygen-containing aromatic plastic wastes, from which organic hydrogen carriers (LOHCs) can be derived. We developed a catalytic system comprised of Ru-ReOx /SiO2 +HZSM-5 for direct HDO of polycarbonate (PC), polyethylene terephthalate (PET), polybutylene terephthalate (PBT), polyphenylene oxide (PPO), and their mixtures, to cycloalkanes as LOHCs, with high yields up to 99 %, under mild reaction conditions. The theoretical hydrogen storage capacity reaches ca. 5.74 wt%. The reaction pathway involves depolymerization of PC into C15 aromatics and C15 monophenols by direct hydrogenolysis of the C-O bond between the benzene ring and ester group, and subsequent parallel hydrogenation of C15 aromatics and HDO of C15 monophenols. HDO of cyclic alcohol is the rate-determining step. The active site is Ru metallic nanoparticles with partially covered ReOx species. The excellent performance is attributed to the synergetic effect of oxophilic ReOx species and Ru metallic sites for C-O hydrogenolysis and hydrogenation, and the promotion effect of HZSM-5 for dehydration of cyclic alcohol. The highly efficient and stable dehydrogenation of cycloalkanes over Pt/γ-Al2 O3 confirms that HDO products can act as LOHCs.

Revealing the promotion of carbonyl groups on vacancy stabilized Pt4/nanocarbons for propane dehydrogenation.

Nanocarbons are promising supports for Pt clusters applied in propane dehydrogenation (PDH), owing to their large surface areas and tunable chemical properties. The vacancies and oxygen-containing groups (OCGs) in nanocarbons can enhance catalytic performance by tailoring the coordination environment of Pt clusters. Herein, 46 nanocarbons with coexisting vacancies and OCGs were designed to support Pt clusters, of which the influences on PDH were revealed by density functional theory calculations. Nanocarbons with divacancies (V2) and CO edge groups were screened out as the most appropriate support for Pt clusters in PDH. Due to the V2, tetrahedral Pt clusters were distorted into three-layered configurations, contributing to enhanced binding strength and a favorable reactive pathway starting from the methylene group in propane. This changed the rate-determining step to the first C-H bond scission with a low energy barrier. The introduction of CO edge groups coexisting with V2 further improved the stabilization of Pt clusters, resulting from the increased electron transfer from Pt atoms to C atoms. The abilities to break C-H bonds and inhibit C-C bond cracking were also enhanced as compared to the nanocarbons with only V2. Therefore, this work provides references on the regulation of vacancies and OCGs in carbon-based catalysts.

Decomposition Mechanism of Isoprenoid Hydrocarbon p-Menthane in the Presence of Pt@FGS Nanoparticles: A ReaxFF-MD Study.

Pt@FGS nanoparticles have shown effective enhancement in the decomposition of hydrocarbon fuels. To further explore the potential enhancing mechanisms of Pt@FGS nanoparticles, the catalytic decomposition of p-menthane, a bioderived isoprenoid "drop-in" fuel with great promise, is investigated here using the reactive force-field molecular dynamics (ReaxFF-MD) simulations. The results show that the Pt@FGS nanoparticles exhibit good catalytic reactivity with a reduction of the activation energy by nearly 62%. Possible initial reactions of enhanced p-menthane (PMT) decomposition are discussed, which suggests that the supported Pt-cluster plays a key role in the dehydrogenation of PMT, as does the oxygen-containing functional group of the functionalized graphene sheets (FGS). It is also interesting to note that the presence of Pt@FGS causes the initial reactions, which are dominated by H-abstraction, favorable in both kinetics and thermodynamics.

Tailoring the catalytic performance of single platinum anchored on graphene by vacancy engineering for propane dehydrogenation: a theoretical study.

Propane dehydrogenation (PDH) is an effective approach to produce propylene. Downsizing the Pt species to single atom catalysts (SACs) has become a hotspot, owing to the maximum utilization and excellent catalytic behavior. However, the agglomeration of SACs is the decisive limitation for high temperature PDH. Herein, single Pt atoms were anchored on graphene with different types of vacancies, and their catalytic performances on PDH were explored based on density functional theory (DFT). As the vacancy size increased, the catalytic activity decreased. It was because the combined site of the detached H atom in propane would transfer from the Pt atom to the C atom around vacancies, thus increasing the migration distance and lowering the activity. However, with the increase of vacancy size, the selectivity to propylene was improved, owing to the enhanced repulsion between C atoms in graphene and propylene. Therefore, instead of stabilizing the single atom, vacancies in carbon materials can also tailor the catalytic performance by geometric disturbance. This fundamental work opens up the possibility for purposeful SAC design in PDH.

Template Guiding for the Encapsulation of Uniformly Subnanometric Platinum Clusters in Beta-Zeolites Enabling High Catalytic Activity and Stability.

Subnanometric metal clusters have attracted extensive attention because of their unique properties as heterogeneous catalysts. However, it is challenging to obtain uniformly distributed metal clusters under synthesis and reaction conditions. Herein, we report a template-guidance protocol to synthesize subnanometric metal clusters uniformly encapsulated in beta-zeolite, with the metal ions anchored to the internal channels of the zeolite template via electrostatic interactions. Pt metal clusters with a narrow size range of 0.89 to 1.22 nm have been obtained on the intersectional sites of beta-zeolite (Pt@beta) with a broad range of Si/Al molar ratios (15-200). The uniformly distributed Pt clusters in Pt@H-beta are subject to strong electron withdrawal by the zeolite, which promotes transfer of active hydrogen, providing excellent activity and stability in hydrodeoxygenation reactions. A general strategy is thus proposed for the encapsulation of subnanometric metal clusters in zeolites with high thermal stability.

Direct Measurement of Radical-Catalyzed C6H6 Formation from Acetylene and Validation of Theoretical Rate Coefficients for C2H3 + C2H2 and C4H5 + C2H2 Reactions.

The addition of vinylic radicals to acetylene is an important step contributing to the formation of polycyclic aromatic hydrocarbons in combustion. The overall reaction 3C2H2 → C6H6 could result in large benzene yields, but without accurate rate parameters validated by experiment, the extent of aromatic ring formation from this pathway is uncertain. The addition of vinyl radicals to acetylene was investigated using time-resolved photoionization time-of-flight mass spectrometry at 500 and 700 K and 5-50 Torr. The formation of C6H6 was observed at all conditions, attributed to sequential addition to acetylene followed by cyclization. Vinylacetylene (C4H4) was observed with increasing yield from 500 to 700 K, attributed to the ß-scission of the thermalized 1,3-butadien-1-yl radical and the chemically activated reaction C2H3 + C2H2 → C4H4 + H. The measured kinetics and product distributions are consistent with a kinetic model constructed using pressure- and temperature-dependent reaction rate coefficients computed from previously reported ab initio calculations. The experiments provide direct measurements of the hypothesized C4H5 intermediates and validate predictions of pressure-dependent addition reactions of vinylic radicals to C2H2, which are thought to play a key role in soot formation.

MiR-6838-5p suppresses cell metastasis and the EMT process in triple-negative breast cancer by targeting WNT3A to inhibit the Wnt pathway.

BACKGROUND: Triple-negative breast cancer (TNBC) has become a common tumor that harms women's physical and mental health, as characterized by a relatively rapid recurrence and a high incidence of brain metastasis. Research increasingly suggests that microRNAs play key roles in the progress of TNBC. However, the function of miR-6838-5p in TNBC has not yet been reported, and requires additional exploration. METHODS: In the present study, we uncovered miR-6838-5p expression in TNBC cells via a quantitative reverse transcriptase-polymerase chain reaction. Functionally, the impacts of up-regulated or down-regulated miR-6838-5p on TNBC invasiveness, Wnt pathway activation and epithelial-mesenchymal transition (EMT) were investigated via transwell and western blot assays. Mechanical analyses were utilized to unmask the miR-6838-5p mechanism in TNBC, including luciferase reporter, western blot and RIP assays. Rescue assays manifested the miR-6838-5p/WNT3A network in TNBC invasiveness through the Wnt pathway. RESULTS: Under-expressed miR-6838-5p was found in TNBC cells. Up-regulation of miR-6838-5p suppressed TNBC cell invasion, migration and blockade of the Wnt pathway. However, down-regulation of miR-6838-5p led to opposite results. Furthermore, we found, via luciferase reporter, western blot and RIP assays, that miR-6838-5p could bind with WNT3A and negatively regulate WNT3A expression. Through rescue experiments, we demonstrated that the overexpression of WNT3A partially rescued the miR-6838-5p overexpression-mediated inhibitory effect, and knockdown of WNT3A partially rescued the miR-6838-5p suppression-mediated promotive effect on the progression of TNBC. CONCLUSIONS: In summary, the results of the present study indicate that miR-6838-5p suppresses cell proliferation, metastasis and the EMT process in TNBC by targeting WNT3A to inhibit the Wnt pathway, which may provide a new insight into the therapeutic strategies of TNBC.

Biological and Chemical Diversity of Bacteria Associated with a Marine Flatworm.

The aim of this research is to explore the biological and chemical diversity of bacteria associated with a marine flatworm Paraplanocera sp., and to discover the bioactive metabolites from culturable strains. A total of 141 strains of bacteria including 45 strains of actinomycetes and 96 strains of other bacteria were isolated, identified and fermented on a small scale. Bioactive screening (antibacterial and cytotoxic activities) and chemical screening (ultra-performance liquid chromatography-mass spectrometry (UPLC-MS)) yielded several target bacterial strains. Among these strains, the ethyl acetate (EA) crude extract of Streptomyces sp. XY-FW47 fermentation broth showed strong antibacterial activity against methicillin-resistant Staphylococcus aureus ATCC43300 (MRSA ATCC43300) and potent cytotoxic effects on HeLa cells. The UPLC-MS spectral analysis of the crude extract indicated that the strain XY-FW47 could produce a series of geldanamycins (GMs). One new geldanamycin (GM) analog, 4,5-dihydro-17-O-demethylgeldanamycin (1), and three known GMs (2-4) were obtained. All of these compounds were tested for antibacterial, cytotoxic, and antifungal activities, yet only GM (3) showed potent cytotoxic (HeLa cells, EC50 = 1.12 µg/mL) and antifungal (Setosphaeria turcica MIC = 2.40 µg/mL) activities. Their structure-activity relationship (SAR) was also preliminarily discussed in this study.

A Secure and Privacy-Preserving Navigation Scheme Using Spatial Crowdsourcing in Fog-Based VANETs.

Fog-based VANETs (Vehicular ad hoc networks) is a new paradigm of vehicular ad hoc networks with the advantages of both vehicular cloud and fog computing. Real-time navigation schemes based on fog-based VANETs can promote the scheme performance efficiently. In this paper, we propose a secure and privacy-preserving navigation scheme by using vehicular spatial crowdsourcing based on fog-based VANETs. Fog nodes are used to generate and release the crowdsourcing tasks, and cooperatively find the optimal route according to the real-time traffic information collected by vehicles in their coverage areas. Meanwhile, the vehicle performing the crowdsourcing task can get a reasonable reward. The querying vehicle can retrieve the navigation results from each fog node successively when entering its coverage area, and follow the optimal route to the next fog node until it reaches the desired destination. Our scheme fulfills the security and privacy requirements of authentication, confidentiality and conditional privacy preservation. Some cryptographic primitives, including the Elgamal encryption algorithm, AES, randomized anonymous credentials and group signatures, are adopted to achieve this goal. Finally, we analyze the security and the efficiency of the proposed scheme.

Quantitative determination of 15 bioactive triterpenoid saponins in different parts of Acanthopanax henryi by HPLC with charged aerosol detection and confirmation by LC-ESI-TOF-MS.

Triterpenoid saponins are difficult to analyze using high-performance liquid chromatography coupled to UV/vis spectrophotometry due to their lack of chromophores. This study describes the first analytical method for the determination of 15 triterpenoid saponins from the leaves, stems, root bark, and fruits of Acanthopanax henryi, using a high-performance liquid chromatography with charged aerosol detection coupled with electrospray ionization mass spectrometry method. The separation was carried out on a Kinetex XB-C18 column with an acetonitrile/water gradient as the mobile phase, followed by charged aerosol detection. The operating conditions of charged aerosol detection were set at 24 kPa for nitrogen pressure and 100 pA for the detection range. Liquid chromatography with electrospray ionization mass spectrometry is described for the identification of compounds in plant samples. The electrospray ionization mass spectrometry method involved the use of the [M + Na](+) and [M + NH4 ](+) ions for compounds 1-15 in the positive ion mode with an extracted ion chromatogram. The developed method was fully validated in terms of linearity, sensitivity, precision, repeatability, and recovery, then subsequently applied to evaluate the quality of A. henryi.

Kinetics and Products of Vinyl + 1,3-Butadiene, a Potential Route to Benzene.

The reaction between vinyl radical, C2H3, and 1,3-butadiene, 1,3-C4H6, has long been recognized as a potential route to benzene, particularly in 1,3-butadiene flames, but the lack of reliable rate coefficients has hindered assessments of its true contribution. Using laser flash photolysis and visible laser absorbance (λ = 423.2 nm), we measured the overall rate coefficient for C2H3 + 1,3-C4H6, k1, at 297 K ≤ T ≤ 494 K and 4 ≤ P ≤ 100 Torr. k1 was in the high-pressure limit in this range and could be fit by the simple Arrhenius expression k1 = (1.1 ± 0.2) × 10(-12) cm(3) molecule(-1) s(-1) exp(-9.9 ± 0.6 kJ mol(-1)/RT). Using photoionization time-of-flight mass spectrometry, we also investigated the products formed. At T ≤ 494 K and P = 25 Torr, we found only C6H9 adduct species, while at 494 K ≤ T ≤ 700 K and P = 4 Torr, we observed ≤∼10% branching to cyclohexadiene in addition to C6H9. Quantum chemistry master-equation calculations using the modified strong collision model indicate that n-C6H9 is the dominant product at low temperature, consistent with our experimental results, and predict the rate coefficient and branching ratios at higher T where chemically activated channels become important. Predictions of k1 are in close agreement with our experimental results, allowing us to recommend the following modified Arrhenius expression in the high-pressure limit from 300 to 2000 K: k1 = 6.5 × 10(-20) cm(3) molecule(-1) s(-1) T(2.40) exp(-1.76 kJ mol(-1)/RT).

Hydrodeoxygenation of Oxygen-Containing Aromatic Plastic Wastes into Cycloalkanes and Aromatics.

Chemical recycling and upcycling offer promising approaches for the management of plastic wastes. Hydrodeoxygenation (HDO) is one of the appealing ways for conversion of oxygen-containing plastic wastes, including polyethylene terephthalate (PET), polybutylene terephthalate (PBT), polycarbonate (PC), polyphenyl ether (PPO), and polyether ether ketone (PEEK), into cyclic alkanes and aromatics in high yields under mild reaction conditions. The challenge lies in achieving C-O activation while preserving C-C bonds. In this review, we highlight the recent advancements in catalytic strategies and catalysts for the conversion of these oxygen-containing plastic wastes into cycloalkanes and aromatics. The reaction systems, including multi-step routes, direct HDO and transfer HDO methods, are exemplified. The design and performance of HDO catalysts are systematically summarized and compared. We comprehensively discuss the functions of the catalysts' components, reaction pathway and mechanism to gain insights into the HDO process for efficient valorization of oxygen-containing plastic wastes. Finally, we provide perspectives for this field, with specific emphasis on the non-noble metal catalyst design, selectivity control, reaction network and mechanism studies, mixed plastic wastes management and product functionalization. We anticipate that this review will inspire innovations on the catalytic process development and rational catalyst design for the HDO of oxygen-containing aromatic plastics to establish a low-emission circular economy.

Crystal Facet Structure Dependence and Promising Pd-Pt Catalytic Materials for Perhydroacenaphthene Dehydrogenation.

Designing an effective Pd-Pt catalytic material with excellent catalytic performance for perhydroacenaphthene (PHAN) dehydrogenation is a great challenge. In this work, in order to explore the crystal facet structure over the bimetallic Pd-Pt catalyst on the dehydrogenation performance of PHAN, the surface compositions of two kinds of Pd (Pt) atoms with different coverage on Pd modulated Pt (PdPt) and Pt modulated Pd (PtPd) catalysts were designed and studied by means of density functional theory (DFT). Through the investigation of the reaction path of PHAN dehydrogenation on PdMLPt(111) and PtMLPd(111) surfaces, it was found that PdMLPt(111) was advantageous to PHAN dehydrogenation (Ea = 2.317 eV). This was attributed to a lower energy barrier, more stable dehydrogenation products, and the fact that Pd doping brought Pt(111) close to the Fermi level. Apparently, Pd modulated Pt catalyst has a broad application prospect in the dehydrogenation of PHAN. In the process of optimizing the PdPt morphology, a method for selecting the minimum active unit of PdPt catalysts with different ratios was proposed, that is, the most stable active unit: rhombus structure was determined according to the surface formation energy. Moreover, we correlated the relationship among the number of H atoms removed, adsorption energy, surface charge, activation energy, reaction energy, and surface coverage, and obtained the important parameters to predict the performance of PdPt catalyst in PHAN dehydrogenation system: surface charge and d-band center. Finally, the essential correlativity among Pd-Pt surface characteristics, catalytic PHAN activity, and adsorption energy was constructed; that is, the relationship model among d-band center, H atom, and product C12H8 adsorption energy was established. This work opens a new simultaneous path to improve the catalytic performance of Pd-Pt-based catalytic materials for PHAN dehydrogenation, which can be achieved by regulating the rhombic active units of Pt modulated by Pd.

Selective H2O2 Electrosynthesis over Defective Carbon from Electrochemical Etching of Molybdenum Carbide.

The controllable synthesis of specific defective carbon catalysts is crucial for two-electron oxygen reduction reaction (2e- ORR) to generate H2O2 due to the great potential applications. Herein, the defective carbon catalysts (Mo-CDC-ns) were prepared by an electrochemical activation (ECA) method with Mo2C/C as a parent. Electrochemical cyclic voltammetry curves, X-ray photoelectron spectroscopy, inductively coupled plasma-mass spectrometry, scanning electron microscopy, and high-resolution transmission electron microscopy confirm the evolution process of a defective carbon structure from the Mo2C phase in which Mo species are first oxidized to Mo6+ species and then the latter are dissolved into the solution and defective carbon is simultaneously formed. Raman and electron paramagnetic resonance spectra reveal that the defect types in Mo-CDC-ns are the edge defect and vacancy defect sites. Compared with the parent Mo2C/C, Mo-CDC-ns exhibit gradually increased kinetic current density and selectivity for H2O2 generation with an extension of activation cycles from 10 (Mo-CDC-10) to 30 (Mo-CDC-30). Over Mo-CDC-30, a kinetic current density of 19.4 mA cm-2 and a selectivity close to 90% in 0.1 M KOH solution were achieved, as well as good stability for H2O2 production in an extended test up to 12 h in an H-cell. Graphene planes and Stone Wales 5757-carbon were constructed as basic models for density functional theory calculations. It revealed that the obtained defective structure after the removal of Mo atoms contains the double vacancy at the edge of graphene (Edge-DVC) and the topological defect on the plane of 5757-carbon (5757C-D), which show more moderate reaction free energy for forming *OOH and smaller energy barrier of 2e- ORR.

A preliminary study of the salivary microbiota of young male subjects before, during, and after acute high-altitude exposure.

Background: The microbial community structure in saliva differs at different altitudes. However, the impact of acute high-altitude exposure on the oral microbiota is unclear. This study explored the impact of acute high-altitude exposure on the salivary microbiome to establish a foundation for the future prevention of oral diseases. Methods. Unstimulated whole saliva samples were collected from 12 male subjects at the following three time points: one day before entering high altitude (an altitude of 350 m, pre-altitude group), seven days after arrival at high altitude (an altitude of 4,500 m, altitude group) and seven days after returning to low altitude (an altitude of 350 m, post-altitude group). Thus, a total of 36 saliva samples were obtained. 16S rRNA V3-V4 region amplicon sequencing was used to analyze the diversity and structure of the salivary microbial communities, and a network analysis was employed to investigate the relationships among salivary microorganisms. The function of these microorganisms was predicted with a Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) analysis. Results: In total, there were 756 operational taxonomic units (OTUs) identified, with 541, 613, and 615 OTUs identified in the pre-altitude, altitude, and post-altitude groups, respectively. Acute high-altitude exposure decreased the diversity of the salivary microbiome. Prior to acute high-altitude exposure, the microbiome mainly consisted of Proteobacteria, Firmicutes, Bacteroidetes, Fusobacteria, and Actinobacteria. After altitude exposure, the relative abundance of Streptococcus and Veillonella increased, and the relative abundance of Prevotella, Porphyromonas, and Alloprevotella decreased. The relationship among the salivary microorganisms was also affected by acute high-altitude exposure. The relative abundance of carbohydrate metabolism gene functions was upregulated, while the relative abundance of coenzyme and vitamin metabolism gene functions was downregulated. Conclusion: Rapid high-altitude exposure decreased the biodiversity of the salivary microbiome, changing the community structure, symbiotic relationships among species, and abundance of functional genes. This suggests that the stress of acute high-altitude exposure influenced the stability of the salivary microbiome.

HGR-Net: Hierarchical Graph Reasoning Network for Arbitrary Shape Scene Text Detection.

As a prerequisite step of scene text reading, scene text detection is known as a challenging task due to natural scene text diversity and variability. Most existing methods either adopt bottom-up sub-text component extraction or focus on top-down text contour regression. From a hybrid perspective, we explore hierarchical text instance-level and component-level representation for arbitrarily-shaped scene text detection. In this work, we propose a novel Hierarchical Graph Reasoning Network (HGR-Net), which consists of a Text Feature Extraction Network (TFEN) and a Text Relation Learner Network (TRLN). TFEN adaptively learns multi-grained text candidates based on shared convolutional feature maps, including instance-level text contours and component-level quadrangles. In TRLN, an inter-text graph is constructed to explore global contextual information with position-awareness between text instances, and an intra-text graph is designed to estimate geometric attributes for establishing component-level linkages. Next, we bridge the cross-feed interaction between instance-level and component-level, and it further achieves hierarchical relational reasoning by learning complementary graph embeddings across levels. Experiments conducted on three publicly available benchmarks SCUT-CTW1500, Total-Text, and ICDAR15 have demonstrated that HGR-Net achieves state-of-the-art performance on arbitrary orientation and arbitrary shape scene text detection.