FMD-UNet: fine-grained feature squeeze and multiscale cascade dilated semantic aggregation dual-decoder UNet for COVID-19 lung infection segmentation from CT images.

With the advancement of computer-aided diagnosis, the automatic segmentation of COVID-19 infection areas holds great promise for assisting in the timely diagnosis and recovery of patients in clinical practice. Currently, methods relying on U-Net face challenges in effectively utilizing fine-grained semantic information from input images and bridging the semantic gap between the encoder and decoder. To address these issues, we propose an FMD-UNet dual-decoder U-Net network for COVID-19 infection segmentation, which integrates a Fine-grained Feature Squeezing (FGFS) decoder and a Multi-scale Dilated Semantic Aggregation (MDSA) decoder. The FGFS decoder produces fine feature maps through the compression of fine-grained features and a weighted attention mechanism, guiding the model to capture detailed semantic information. The MDSA decoder consists of three hierarchical MDSA modules designed for different stages of input information. These modules progressively fuse different scales of dilated convolutions to process the shallow and deep semantic information from the encoder, and use the extracted feature information to bridge the semantic gaps at various stages, this design captures extensive contextual information while decoding and predicting segmentation, thereby suppressing the increase in model parameters. To better validate the robustness and generalizability of the FMD-UNet, we conducted comprehensive performance evaluations and ablation experiments on three public datasets, and achieved leading Dice Similarity Coefficient (DSC) scores of 84.76, 78.56 and 61.99% in COVID-19 infection segmentation, respectively. Compared to previous methods, the FMD-UNet has fewer parameters and shorter inference time, which also demonstrates its competitiveness.

Integration analysis of miRNA-mRNA uncovers the molecular immune mechanism of macrophage response to Aeromonas veronii infection.

Macrophages are innate immunity cells which play pivotal roles in infectious immunity. Aeromonas veronii is a zoonotic agent capable of causing sepsis and poses a serious threat to public health. However, few studies have focused on miRNA-mRNA integration analysis to address the immune mechanisms of macrophage response to A. veronii infection. Herein, we characterized the immunophysiological, biochemical, and transcriptome changes of macrophage under A. veronii infection. We found that macrophages infected with A. veronii released large amounts of cytokines and triggered NLRP3-dependent pyroptosis. Subsequently, 603 differentially expressed miRNAs (DEMIs) and 3693 differentially expressed mRNAs (DEMs) were identified by RNA-seq analysis under A. veronii infection. Moreover, integrated analysis of miRNA-mRNA yielded 66 miRNA-target gene pairs composed of 41 DEMIs and 27 DEMs. We next identified the Toll-like receptor, NOD-like receptor, TNF and NF-κB pathways as necessary for macrophage to respond to A. veronii infection. miR-847 and miR-627 were involved in macrophage response to A. veronii infection by negatively regulating Pannexin-1 and thioredoxin interacting protein (TXNIP). Our findings elucidate the molecular mechanism of macrophage response to A. veronii infection at the miRNA level, providing many candidate miRNAs and mRNAs therapeutic targets for the prevention and treatment of A. veornii infectious diseases.

Residual and bidirectional LSTM for epileptic seizure detection.

Electroencephalogram (EEG) plays a pivotal role in the detection and analysis of epileptic seizures, which affects over 70 million people in the world. Nonetheless, the visual interpretation of EEG signals for epilepsy detection is laborious and time-consuming. To tackle this open challenge, we introduce a straightforward yet efficient hybrid deep learning approach, named ResBiLSTM, for detecting epileptic seizures using EEG signals. Firstly, a one-dimensional residual neural network (ResNet) is tailored to adeptly extract the local spatial features of EEG signals. Subsequently, the acquired features are input into a bidirectional long short-term memory (BiLSTM) layer to model temporal dependencies. These output features are further processed through two fully connected layers to achieve the final epileptic seizure detection. The performance of ResBiLSTM is assessed on the epileptic seizure datasets provided by the University of Bonn and Temple University Hospital (TUH). The ResBiLSTM model achieves epileptic seizure detection accuracy rates of 98.88-100% in binary and ternary classifications on the Bonn dataset. Experimental outcomes for seizure recognition across seven epilepsy seizure types on the TUH seizure corpus (TUSZ) dataset indicate that the ResBiLSTM model attains a classification accuracy of 95.03% and a weighted F1 score of 95.03% with 10-fold cross-validation. These findings illustrate that ResBiLSTM outperforms several recent deep learning state-of-the-art approaches.

Uncovering the Interaction between TRAF1 and MAVS in the RIG-I Pathway to Enhance the Upregulation of IRF1/ISG15 during Classical Swine Fever Virus Infection.

Classical swine fever (CSF) is caused by the classical swine fever virus (CSFV), which poses a threat to swine production. The activation of host innate immunity through linker proteins such as tumor necrosis factor receptor (TNF-R)-associated factor (TRAF) is crucial for the induction of the NF-κB pathway. Recent research has revealed the involvement of mitochondrial antiviral-signaling protein (MAVS) in the interaction with TRAF2, 3, 5, and 6 to activate both the NF-κB and IRF3 pathways. This study revealed that CSFV infection led to the upregulation of TRAF1 mRNA and protein levels; moreover, TRAF1 overexpression inhibited CSFV replication, while TRAF1 knockdown promoted replication, highlighting its importance in the host response to CSFV infection. Additionally, the expression of RIG-I, MAVS, TRAF1, IRF1, and ISG15 were detected in PK-15 cells infected with CSFV, revealing that TRAF1 plays a role in regulating IRF1 and ISG15 within the RIG-I pathway. Furthermore, Co-IP, GST pull-down, and IFA analyses demonstrated that TRAF1 interacted with MAVS and co-localized in the cytoplasm during CSFV infection. Ultimately, TRAF1 acted as a novel member of the TRAF family, bound to MAVS as a linker molecule, and functioned as a mediator downstream of MAVS in the RIG-I/MAVS pathway against CSFV replication.

An improved approach for automated cervical cell segmentation with PointRend.

Regular screening for cervical cancer is one of the best tools to reduce cancer incidence. Automated cell segmentation in screening is an essential task because it can present better understanding of the characteristics of cervical cells. The main challenge of cell cytoplasm segmentation is that many boundaries in cell clumps are extremely difficult to be identified. This paper proposes a new convolutional neural network based on Mask RCNN and PointRend module, to segment overlapping cervical cells. The PointRend head concatenates fine grained features and coarse features extracted from different feature maps to fine-tune the candidate boundary pixels of cell cytoplasm, which are crucial for precise cell segmentation. The proposed model achieves a 0.97 DSC (Dice Similarity Coefficient), 0.96 TPRp (Pixelwise True Positive Rate), 0.007 FPRp (Pixelwise False Positive Rate) and 0.006 FNRo (Object False Negative Rate) on dataset from ISBI2014. Specially, the proposed method outperforms state-of-the-art result by about 3 % on DSC, 1 % on TPRp and 1.4 % on FNRo respectively. The performance metrics of our model on dataset from ISBI2015 are slight better than the average value of other approaches. Those results indicate that the proposed method could be effective in cytological analysis and then help experts correctly discover cervical cell lesions.

Apolipoprotein D facilitates rabies virus propagation by interacting with G protein and upregulating cholesterol.

Rabies virus (RABV) causes a fatal neurological disease, consisting of unsegmented negative-strand RNA, which encodes five structural proteins (3'-N-P-M-G-L-5'). Apolipoprotein D (ApoD), a lipocalin, is upregulated in the nervous system after injury or pathological changes. Few studies have focused on the role of ApoD during virus infection so far. This study demonstrated that ApoD is upregulated in the mouse brain (in vivo) and C8-D1A cells (in vitro) after RABV infection. By upregulating ApoD expression in C8-D1A cells, we found that ApoD facilitated RABV replication. Additionally, Co-immunoprecipitation demonstrated that ApoD interacted with RABV glycoprotein (G protein). The interaction could promote RABV replication by upregulating the cholesterol level. These findings revealed a novel role of ApoD in promoting RABV replication and provided a potential therapeutic target for rabies.

Tissue factor regulates autophagy in pulmonary artery endothelial cells from chronic thromboembolic pulmonary hypertension rats via the p38 MAPK-FoxO1 pathway.

AIMS: To detect the expression of autophagy components, p38 MAPK (p38) and phosphorylated forkhead box transcription factor O-1 (pFoxO1) in pulmonary vascular endothelial cells of chronic thromboembolic pulmonary hypertension (CTEPH) rats and to investigate the possible mechanism through which tissue factor (TF) regulates autophagy. METHODS: Pulmonary artery endothelial cells (PAECs) were isolated from CTEPH (CTEPH group) and healthy rats (control group (ctrl group)) which were cocultured with TF at different time points including 12 h, 24 h, 48 h and doses including 0 nM,10 nM, 100 nM, 1µM, 10µM, 100µM and cocultured with TFPI at 48 h including 0 nM, 2.5 nM, 5 nM. The expression of forkhead box transcription factor O-1 (FoxO1), pFoxO1, p38, Beclin-1 and LC3B in PAECs was measured. Coimmunoprecipitation (co-IP) assays were used to detect the interaction between FoxO1 and LC3. RESULTS: The protein expression of p-FoxO1/FoxO1 was significantly lower in the CTEPH groups (cocultured with TF from 0 nM to 100 µM) than in the ctrl group at 12 h, 24 h, and 48 h (P < 0.05) and was significantly lower in the CTEPH groups (cocultured with TFPI from 0 nM to 5 nM) than in the ctrl group at 48 h (P < 0.05). The protein expression of p38 in the CTEPH groups treated with 0 nM, 10 nM, 100 nM or 1 µM TF for 48 h significantly increased than ctrl groups (P < 0.05) and was significantly increased in the CTEPH groups (cocultured with TFPI concentration from 0 nM to 5 nM) than in the ctrl group at 48 h (P < 0.05). The protein expression of Beclin1 at the same concentration (cocultured with TF from 0 nM to 100 µM) was significantly lower in the CTEPH groups than ctrl groups after 24 h and 48 h (P < 0.05) and was significantly decreased in the CTEPH groups (cocultured with TFPI concentration from 2.5 nM to 5 nM) than in the ctrl group at 48 h (P < 0.05). The protein expression of LC3-II/LC3-I at the same concentration (cocultured with TF 0 nM, 1 µM, 10 µM, and 100 µM) was significantly lower in the CTEPH than in the ctrl groups after 12 h (P < 0.05) and was significantly lower in the CTEPH groups (cocultured with TFPI concentration from 0 nM to 5 nM) than in the ctrl group at 48 h (P < 0.05). There were close interactions between FoxO1 and LC3 in the control and CTEPH groups at different doses and time points. CONCLUSION: The autophagic activity of PAECs from CTEPH rats was disrupted. TF, FoxO1 and p38 MAPK play key roles in the autophagic activity of PAECs. TF may regulate autophagic activity through the p38 MAPK-FoxO1 pathway.

A Waterproof Flexible Paper-Based Thermoelectric Generator for Humidity and Underwater Environments.

A thermoelectric generator (TEG) is one of the important energy harvesting sources for wearable electronic devices, which converts waste heat into electrical energy without any external stimuli, such as light or mechanical motion. However, the poor flexibility of traditional TEGs (e.g., Si-based TE devices) causes the limitations in practical applications. Flexible paper substrates are becoming increasingly attractive in wearable electronic technology owing to their usability, environmental friendliness (disposable, biodegradable, and renewable materials), and foldability. The high water-absorbing quality of paper restricts its scope of application due to water failure. Therefore, we propose a high-performance flexible waterproof paper-based thermoelectric generator (WPTEG). A modification method that infiltrates TE materials into cellulose paper through vacuum filtration is used to prepare the TE modules. By connecting the TE-modified paper with Al tape, as well as a superhydrophobic layer encapsulation, the WPTEG is fabricated. The WPTEG with three P-N modules can generate an output voltage of up to 235 mV at a temperature difference of 50 K, which can provide power to portable electronic devices such as diodes, clocks, and calculators in hot water. With the waterproof property, the WPTEG paves the way for achieving multi-scenario applications in humid environments on human skin.

Crystal structure of methyltransferase CbiL from Akkermansia muciniphila.

Akkermansia muciniphila is a mucin-degrading probiotic that colonizes the gastrointestinal tract. Genomic analysis identified a set of genes involved in the biosynthesis of corrin ring, including the cobalt factor II methyltransferase CbiL, in some phylogroups of A. muciniphila, implying a potential capacity for de novo synthesis of cobalamin. In this work, we determined the crystal structure of CbiL from A. muciniphila at 2.3 Å resolution. AmCbiL exists as a dimer both in solution and in crystal, and each protomer consists of two α/ß domains, the N-terminal domain and the C-terminal domain, consistent with the folding of typical class III MTases. The two domains create an open trough, potentially available to bind the substrates SAM and cobalt factor II. Sequence and structural comparisons with other CbiLs, assisted by computer modeling, suggest that AmCbiL should have cobalt factor II C-20 methyltransferase activity. Our results support that certain strains of A. muciniphila may be capable of synthesizing cobalamin de novo.

A novel exoskeletal-derived C-type lectin facilitates phagocytosis of hemocytes in the oriental river prawn Macrobrachium nipponense.

C-type lectins (CTLs) execute critical functions in multiple immune responses of crustaceans as a member of pattern recognition receptors (PRRs) family. In this study, a novel CTL was identified from the exoskeleton of the oriental river prawn Macrobrachium nipponense (MnLec3). The full-length cDNA of MnLec3 was 1150 bp with an open reading frame of 723 bp, encoding 240 amino acids. MnLec3 protein contained a signal peptide and one single carbohydrate-recognition domain (CRD). MnLec3 transcripts were widely distributed at the exoskeleton all over the body. Significant up-regulation of MnLec3 in exoskeleton after Aeromonas hydrophila challenged suggested the involvement of MnLec3 as well as the possible function of the exoskeleton in immune response. In vitro tests with recombinant MnLec3 protein (rMnLec3) manifested that it had polysaccharide binding activity, a wide spectrum of bacterial binding activity and agglutination activity only for tested Gram-negative bacteria (Escherichia coli, Vibrio anguillarum and A. hydrophila). Moreover, rMnLec3 significantly promoted phagocytic ability of hemocytes against A. hydrophila in vivo. What's more, MnLec3 interference remarkably impaired the survivability of the prawns when infected with A. hydrophila. Collectively, these results ascertained that MnLec3 derived from exoskeleton took an essential part in immune defense of the prawns against invading bacteria as a PRR.

Integration Analysis of Single-Cell Multi-Omics Reveals Prostate Cancer Heterogeneity.

Prostate cancer (PCa) is an extensive heterogeneous disease with a complex cellular ecosystem in the tumor microenvironment (TME). However, the manner in which heterogeneity is shaped by tumors and stromal cells, or vice versa, remains poorly understood. In this study, single-cell RNA sequencing, spatial transcriptomics, and bulk ATAC-sequence are integrated from a series of patients with PCa and healthy controls. A stemness subset of club cells marked with SOX9highARlow expression is identified, which is markedly enriched after neoadjuvant androgen-deprivation therapy (ADT). Furthermore, a subset of CD8+CXCR6+ T cells that function as effector T cells is markedly reduced in patients with malignant PCa. For spatial transcriptome analysis, machine learning and computational intelligence are comprehensively utilized to identify the cellular diversity of prostate cancer cells and cell-cell communication in situ. Macrophage and neutrophil state transitions along the trajectory of cancer progression are also examined. Finally, the immunosuppressive microenvironment in advanced PCa is found to be associated with the infiltration of regulatory T cells (Tregs), potentially induced by an FAP+ fibroblast subset. In summary, the cellular heterogeneity is delineated in the stage-specific PCa microenvironment at single-cell resolution, uncovering their reciprocal crosstalk with disease progression, which can be helpful in promoting PCa diagnosis and therapy.

Release Characteristics of Harmful Trace Elements during Dynamic Leaching and Static Immersion of Coal Gangue in Xinjiang.

Coal gangue has dual attributes of waste residue and resources. Clarifying the release characteristics of harmful trace elements from the coal gangue can provide a theoretical basis for environmental impact and resource utilization. In this study, the characteristics of harmful trace elements released from coal gangue in Xinjiang during dynamic leaching and static immersion experiments were determined using proximate analysis, X-ray powder diffraction (XRD), X-ray fluorescence spectrometry (XRF), and inductively coupled plasma mass spectrometry (ICP-MS). The results show that (1) the higher the content of harmful trace elements in coal gangue and the greater the concentration coefficient (CC), the greater the release of elements in dynamic leaching and static immersion experiments. The mode of occurrence of trace elements in the coal gangue determines their transport and release. Elements are associated not only with moisture but also with minerals, such as clays, sulfides, and carbonates, which are readily soluble in water. (2) The release of harmful trace elements was inversely proportional to time in the dynamic leaching experiments, and the main reason for the reduction in element release during the late leaching period was the adsorption effect of clay minerals. In the dynamic leaching experiment, harmful trace elements in the surrounding environment continued to accumulate, and static immersion experiments in water showed that harmful trace elements gradually reached dynamic equilibrium. The concentration of most elements in the late stage of the static immersion experiment was lower than that in the early stage, indicating that the environmental hazards of dynamic leaching were greater than those of the static immersion of coal gangue in Xinjiang.

The impacts of microplastics on the cycling of carbon and nitrogen in terrestrial soil ecosystems: Progress and prospects.

As contaminants of emerging concern, microplastics (MPs) are ubiquitously present in almost all environmental compartments of the earth, with terrestrial soil ecosystems as the major sink for these contaminants. The accumulation of MPs in the soil can trigger a wide range of effects on soil physical, chemical, and microbial properties, which may in turn cause alterations in the biogeochemical processes of some key elements, such as carbon and nitrogen. Until recently, the effects of MPs on the cycling of carbon and nitrogen in terrestrial soil ecosystems have yet to be fully understood, which necessitates a review to summarize the current research progress and propose suggestions for future studies. The presence of MPs can affect the contents and forms of soil carbon and nitrogen nutrients (e.g., total and dissolved organic carbon, dissolved organic nitrogen, NH4+-N, and NO3--N) and the emissions of CH4, CO2, and N2O by altering soil microbial communities, functional gene expressions, and enzyme activities. Exposure to MPs can also affect plant growth and physiological processes, consequently influencing carbon fixation and nitrogen uptake. Specific effects of MPs on carbon and nitrogen cycling and the associated microbial parameters can vary considerably with MP properties (e.g., dose, polymer type, size, shape, and aging status) and soil types, while the mechanisms of interaction between MPs and soil microbes remain unclear. More comprehensive studies are needed to narrow the current knowledge gaps.

Reversible Insertion of [AlCl4 ]- Superhalides in Graphite.

Graphite-based dual-ion batteries are with potential higher energy density, making them a unique candidate in energy storage systems. However, anion insertion into graphite in aqueous environment remains a significant challenge. Herein, we report that the reversible insertion of Al-Cl superhalide into expanded graphite (EG) delivers an ultrahigh specific capacity of ~171 mAh g-1 from an aqueous deep eutectic solvent (DES) gel electrolyte of 50 m ChCl+5 m AlCl3 . High-resolution transmission electron microscopy (HRTEM), Raman spectra and X-ray diffraction (XRD) show that the EG generates turbostratic structure during Al-Cl superhalide (de)insertion instead of presenting typical graphite intercalation compounds (GIC), thus attributing to the high capacity during Al-Cl superhalide insertion.

Multi-behavioral recommendation model based on dual neural networks and contrast learning.

In order to capture the complex dependencies between users and items in a recommender system and to alleviate the smoothing problem caused by the aggregation of multi-layer neighborhood information, a multi-behavior recommendation model (DNCLR) based on dual neural networks and contrast learning is proposed. In this paper, the complex dependencies between behaviors are divided into feature correlation and temporal correlation. First, we set up a personalized behavior vector for users and use a graph-convolution network to learn the features of users and items under different behaviors, and we then combine the features of self-attention mechanism to learn the correlation between behaviors. The multi-behavior interaction sequence of the user is input into the recurrent neural network, and the temporal correlation between the behaviors is captured by combining the attention mechanism. The contrast learning is introduced based on the double neural network. In the graph convolution network layer, the distances between users and similar users and between users and their preference items are shortened, and the distance between users and their short-term preference is shortened in the circular neural network layer. Finally, the personalized behavior vector is integrated into the prediction layer to obtain more accurate user, behavior and item characteristics. Compared with the sub-optimal model, the HR@10 on Yelp, ML20M and Tmall real datasets are improved by 2.5%, 0.3% and 4%, respectively. The experimental results show that the proposed model can effectively improve the recommendation accuracy compared with the existing methods.

Dmbg-Net: Dilated multiresidual boundary guidance network for COVID-19 infection segmentation.

Accurate segmentation of infected regions in lung computed tomography (CT) images is essential for the detection and diagnosis of coronavirus disease 2019 (COVID-19). However, lung lesion segmentation has some challenges, such as obscure boundaries, low contrast and scattered infection areas. In this paper, the dilated multiresidual boundary guidance network (Dmbg-Net) is proposed for COVID-19 infection segmentation in CT images of the lungs. This method focuses on semantic relationship modelling and boundary detail guidance. First, to effectively minimize the loss of significant features, a dilated residual block is substituted for a convolutional operation, and dilated convolutions are employed to expand the receptive field of the convolution kernel. Second, an edge-attention guidance preservation block is designed to incorporate boundary guidance of low-level features into feature integration, which is conducive to extracting the boundaries of the region of interest. Third, the various depths of features are used to generate the final prediction, and the utilization of a progressive multi-scale supervision strategy facilitates enhanced representations and highly accurate saliency maps. The proposed method is used to analyze COVID-19 datasets, and the experimental results reveal that the proposed method has a Dice similarity coefficient of 85.6% and a sensitivity of 84.2%. Extensive experimental results and ablation studies have shown the effectiveness of Dmbg-Net. Therefore, the proposed method has a potential application in the detection, labeling and segmentation of other lesion areas.

A novel "prime and pull" strategy mediated by the combination of two dendritic cell-targeting designs induced protective lung tissue-resident memory T cells against H1N1 influenza virus challenge.

Vaccination is still the most promising strategy for combating influenza virus pandemics. However, the highly variable characteristics of influenza virus make it difficult to develop antibody-based universal vaccines, until now. Lung tissue-resident memory T cells (TRM), which actively survey tissues for signs of infection and react rapidly to eliminate infected cells without the need for a systemic immune reaction, have recently drawn increasing attention towards the development of a universal influenza vaccine. We previously designed a sequential immunization strategy based on orally administered Salmonella vectored vaccine candidates. To further improve our vaccine design, in this study, we used two different dendritic cell (DC)-targeting strategies, including a single chain variable fragment (scFv) targeting the surface marker DC-CD11c and DC targeting peptide 3 (DCpep3). Oral immunization with Salmonella harboring plasmid pYL230 (S230), which displayed scFv-CD11c on the bacterial surface, induced dramatic production of spleen effector memory T cells (TEM). On the other hand, intranasal boost immunization using purified DCpep3-decorated 3M2e-ferritin nanoparticles in mice orally immunized twice with S230 (S230inDC) significantly stimulated the differentiation of lung CD11b+ DCs, increased intracellular IL-17 production in lung CD4+ T cells and elevated chemokine production in lung sections, such as CXCL13 and CXCL15, as determined by RNAseq and qRT‒PCR assays, resulting in significantly increased percentages of lung TRMs, which could provide efficient protection against influenza virus challenge. The dual DC targeting strategy, together with the sequential immunization approach described in this study, provides us with a novel "prime and pull" strategy for addressing the production of protective TRM cells in vaccine design.

[Metabolomic study on urine of chronic inflammation rats treated with Buyang Huanwu Decoction based on UPLC-Q-TOF-MS].

The study investigated the effect of Buyang Huanwu Decoction(BYHWD) on endogenous biomarkers in the urine of rats with chronic inflammation induced by lipopolysaccharide(LPS) using ultra-high performance liquid chromatography-quadrupole-time-of-flight-mass spectrometry(UPLC-Q-TOF-MS), aiming to elucidate the molecular mechanism underlying the therapeutic effect of BYHWD on chronic inflammation from a metabolomics perspective. Male SD rats were randomly divided into a normal group, a model group, and low-, medium-, and high-dose BYHWD groups(7.5, 15, and 30 g·kg~(-1)). The model group and BYHWD groups received tail intravenous injection of LPS(200 µg·kg~(-1)) on the first day of each week, followed by oral administration of BYHWD once a day for four consecutive weeks. Urine samples were collected at the end of the administration period, and UPLC-Q-TOF-MS was used to analyze the metabolic profiles of the rat urine in each group. Multivariate statistical analysis methods such as principal component analysis(PCA), partial least squares-discriminant analysis(PLS-DA), and orthogonal partial least squares-discriminant analysis(OPLS-DA) were used to analyze the effect of BYHWD on endogenous metabolites. One-way ANOVA and variable importance for the projection(VIP) were used to screen for potential biomarkers related to chronic inflammation. The identified biomarkers were subjected to pathway and enrichment analysis using MetaboAnalyst 5.0. A total of 25 potential biomarkers were screened and identified in the rat urine in this experiment. Compared with the normal group, the model group showed significant increases in the levels of 14 substances(P<0.05) and significant decreases in the levels of 11 substances(P<0.05). BYHWD was able to effectively reverse the trend of most endogenous biomarkers. Compared with the model group, BYHWD significantly down-regulated 13 biomarkers(P<0.05) and up-regulated 10 biomarkers(P<0.05). The metabolic products were mainly related to the biosynthesis of pantothenic acid and coenzyme A, tryptophan metabolism, retinol metabolism, and propionate metabolism. BYHWD has therapeutic effect on chronic inflammation induced by LPS, which may be related to its ability to improve the levels of endogenous metabolites, enhance the body's anti-inflammatory and antioxidant capabilities, and restore normal metabolic activity.

Recommendation model based on intention decomposition and heterogeneous information fusion.

In order to solve the problem of timeliness of user and item interaction intention and the noise caused by heterogeneous information fusion, a recommendation model based on intention decomposition and heterogeneous information fusion (IDHIF) is proposed. First, the intention of the recently interacting items and the users of the recently interacting candidate items is decomposed, and the short feature representation of users and items is mined through long-short term memory and attention mechanism. Then, based on the method of heterogeneous information fusion, the interactive features of users and items are mined on the user-item interaction graph, the social features of users are mined on the social graph, and the content features of the item are mined on the knowledge graph. Different feature vectors are projected into the same feature space through heterogeneous information fusion, and the long feature representation of users and items is obtained through splicing and multi-layer perceptron. The final representation of users and items is obtained by combining short feature representation and long feature representation. Compared with the baseline model, the AUC on the Last.FM and Movielens-1M datasets increased by 1.83 and 4.03 percentage points, respectively, the F1 increased by 1.28 and 1.58 percentage points, and the Recall@20 increased by 3.96 and 2.90 percentage points. The model proposed in this paper can better model the features of users and items, thus enriching the vector representation of users and items, and improving the recommendation efficiency.

A Social Recommendation Model Based on Basic Spatial Mapping and Bilateral Generative Adversarial Networks.

Social recommender systems are expected to improve recommendation quality by incorporating social information when there is little user-item interaction data. Therefore, how to effectively fuse interaction information and social information becomes a hot research topic in social recommendation, and how to mine and exploit the heterogeneous information in the interaction and social space becomes the key to improving recommendation performance. In this paper, we propose a social recommendation model based on basic spatial mapping and bilateral generative adversarial networks (MBSGAN). First, we propose to map the base space to the interaction and social space, respectively, in order to overcome the issue of heterogeneous information fusion in two spaces. Then, we construct bilateral generative adversarial networks in both interaction space and social space. Specifically, two generators are used to select candidate samples that are most similar to user feature vectors, and two discriminators are adopted to distinguish candidate samples from high-quality positive and negative examples obtained from popularity sampling, so as to learn complex information in the two spaces. Finally, the effectiveness of the proposed MBSGAN model is verified by comparing it with both eight social recommendation models and six models based on generative adversarial networks on four public datasets, Douban, FilmTrust, Ciao, and Epinions.