Unveiling the molecular mechanism of Qing Wen Bai Du decoction against dengue fever: An integrated study of bioinformatic analysis, machine learning and network pharmacology

Unveiling the molecular mechanism of Qing Wen Bai Du decoction against dengue fever: An integrated study of bioinformatic analysis, machine learning and network pharmacology | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Unveiling the molecular mechanism of Qing Wen Bai Du decoction against dengue fever: An integrated study of bioinformatic analysis, machine learning and network pharmacology Yan Xiao, Zhanchen Liu, Shengjie Hu, Peng Yao, Yajun Liu, Maosheng Cheng This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8520853/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 15 You are reading this latest preprint version Abstract Dengue fever, a mosquito-borne disease caused by the dengue virus (DENV), poses a serious global health challenge with no FDA-approved drug available. Qingwen Baidu Decoction (QBD) is a traditional Chinese medicine prescription used in China to treat critical dengue fever, though its mechanism remains unclear. This study aimed to elucidate the therapeutic mechanisms of QBD against dengue fever through a combination of network pharmacology, machine learning, and molecular docking. Potential targets of QBD were predicted using network pharmacology, and core genes were identified from DENV-upregulated genes via machine learning. Functional enrichment analyses, including GO and KEGG, were conducted to explore related biological processes and pathways. Single-cell RNA sequencing and immune infiltration analyses were performed to identify key cell subtypes and immune correlations. Molecular docking and dynamics simulations were used to evaluate interactions between QBD components and target proteins. Results indicated that QBD interferes with DENV entry by targeting EPHB2 and inhibits viral replication by binding structural protein E and C and nonstructural protein NS5. Additionally, QBD alleviates inflammatory responses by suppressing CXCL10 and EZH2, and modulates host immune responses during DENV infection. Dengue fever Qingwen Baidu Decoction Network pharmacology Bioinformatic analysis Machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Introduction Dengue fever is an acute mosquito-borne infectious disease caused by the dengue virus (DENV). It is primarily transmitted through the bites of Aedes aegypti and Aedes albopictus mosquitoes. The main epidemic areas are tropical and subtropical regions, such as Southeast Asia, the Caribbean, South America and South China. The world health organization (WHO) estimates that about half of the world's population is now at risk of dengue fever with 100–400 million infections occurring each year. The highest number of dengue cases was recorded in 2023, affecting over 80 countries and resulting in over 6.5 million cases and more than 7300 dengue-related deaths. DENV may cause a primary infection or a secondary infection, the latter of which is called as dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS). DHF proves life-threatening due to a wide range of clinical manifestations, including plasma leakage, fluid accumulation, respiratory distress, hemorrhage, and organ damage. Without proper treatment, the mortality rate of DHF may exceed 20%. DENV is a positive-sense single-stranded RNA virus belonging to the Flaviviridae family [ 1 , 2 ]. DENV genome encodes for three structural proteins (C, prM, and E) and seven non-structural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) [ 3 ]. Based on structural differences on their surface, DENV is classified into four distinct serotypes (DENV-1, DENV-2, DENV-3, and DENV-4). Each serotype can cause a spectrum of disease manifestations. Infection with one serotype confers protective immunity against that specific serotype but does not provide cross-protective immunity against other serotypes. Notably, studies have shown that secondary infections with heterologous serotypes are a significant risk factor for developing DHF/DSS [ 4 – 7 ]. Over the past decades, a great effort has been devoted into the research of vaccines and anti-viral drugs against DENV, however there are currently no approved antivirals available to treat dengue fever [ 8 , 9 ]. Dengvaxia is the world's first vaccine for DENV, however it is only approved for use in individuals 6 through 16 years of age with laboratory-confirmed previous dengue infection and living in endemic areas [ 10 ]. Generally, antiviral drug discovery can be approached in two ways, including inhibitors that target viral components and inhibitors that target host cell factors based on the inhibitory mechanism. Some drugs targeting non-structural proteins, such as temoporfin, niclosamide, and nitazoxanide, are currently undergoing clinical trials for the treatment of dengue fever [ 11 , 12 ]. According to the fundamental theory of traditional Chinese medicine, dengue fever belongs to the category of "plague" and "epidemic", which has the characteristics of "dampness, toxicity, stasis and closure". As mentioned in "The Diagnosis and Treatment Guidelines for Dengue Fever" issued by the National Health Commission of China, Qingwen Baidu Decoction (QBD) is used to treat the critical phase of dengue fever [ 13 ]. Originating from the Qing dynasty, QBD is renowned for its effects in heat-clearing and detoxifying, making it one of the most representative antiviral formulas in traditional Chinese medicine [ 14 ]. QBD is composed of 14 traditional Chinese medicinal ingredients: Gypsum Fibrosum, Rhinoceros Unicornis, Rehmannia Glutinosa, Gardeniae Fructus, Coptidis Rhizoma, Platycodon Grandiforus, Anemarrhenae Rhizoma, Scutellariae Radix, Radix Paeoniae Rubra, Moutan Cortex, Figwort Root, Phyllostachys Nigra, Forsythiae Fructus and licorice [ 15 ]. Modern pharmacological studies indicate that QBD is capable of blocking viral entry into cells, inhibiting viral replication, and demonstrating significant suppressive effects against multiple influenza viruses. Moreover, QBD also improves coagulation dysfunction, mitigates organ damage, alleviates clinical symptoms, and enhances patient prognosis [ 16 ]. However, the specific active chemical components responsible for its efficacy against dengue fever, as well as the molecular mechanisms underlying its antiviral effects, remain unclear. Machine learning can provide decision-making support for specific challenges in drug research by analyzing large amounts of bioinformatics data. It has been applied across various stages of drug discovery, including target identification and validation, recognition of prognostic biomarkers, and drug design and screening [ 17 ]. Machine learning aids in identifying potential disease-related molecules and signaling pathways, pinpointing critical therapeutic targets, thereby offering vital insights for novel drug discovery and elucidating the mechanisms of existing drugs [ 18 ]. Our work described here integrates bioinformatics analysis, machine learning and network pharmacology to identify potential therapeutic targets and associated signaling pathways for adjuvant therapeutic pathways of QBD against dengue fever. Subsequently, these potential targets are combined with dengue virus proteins to identify the active chemical ingredients and explain the molecular mechanisms of QBD. The overall workflow of the research design is illustrated in Fig. 1 . Methods Herb-compound-target network construction Natural chemical compounds in the QBD were retrieved from the databases including TCMSP, TCMID and Herb [ 19 – 21 ]. These compounds were initially screened by the following criteria: oral bioavailability (OB > 30%) and drug-likeness (DL > 0.18). Subsequently, SwissADME [ 22 ] was employed to filter compounds adhering to the "rule of five" principles: molecular weight < 500 Da, hydrogen bond donors (HBDs) ≤ 5, hydrogen bond acceptors (HBAs) ≤ 10, lipophilicity (logP) < 5, and rotatable bonds ≤ 10. Potential targets of the obtained compounds were collected from the databases and predicted by the SwissTargetPrediction. The combined targets were used to construct the herb-compound-target network by Cytoscape (version 3.10.1), visualizing interactions between QBD and predicted biological targets. Differential gene expression analysis The DENV-3 dataset (GSE216328) was retrieved from the Gene Expression Omnibus (GEO) database ( https://www.ncbi.nlm.nih.gov/geo/ ). After data preprocessing, including normalization and batch effect correction, differential expression analysis was performed using the R software package DESeq2. Differentially expressed genes (DEGs) were defined using the following criteria: adjusted P 0.5. Volcano plots were generated to visualize the distribution of DEGs based on significance and fold change values, highlighting upregulated (red) and downregulated (blue) genes. Gene Set Enrichment Analysis Gene Set Enrichment Analysis (GSEA) was performed using GSEA v4.3.2 ( https://www.gsea-msigdb.org/gsea/index.jsp ) to evaluate the distribution patterns of DEGs within predefined gene sets, ranked by their phenotypic relevance, to identify their contributions in DENV infection. Gene expression levels were sorted in descending order. The c2.cp.v7.2.symbols.gmt gene set collection was used from the Molecular Signatures Database (MSigDB) to assess enrichment significance. Statistical thresholds were set as false discovery rate (FDR) < 0.25. Enriched pathways were visualized using GSEA-generated enrichment plots and heatmaps. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted on the DEGs using R packages "ggplot2 " and "pathview", with significance defined as adjusted P < 0.05. scRNA-seq analysis Dengue virus single-cell transcriptome data was obtained from the GSE271966 dataset. The single-cell RNA sequencing (scRNA-seq) data was analyzed using the Seurat package in R. Genes expressed in < 3 cells and cells detectable in < 300 genes were removed. Data processing was performed using a single-cell matrix for each sample, and log-normalization was performed to standardize the data for each sample. The FindVariableFeatures function was used to obtain features that vary significantly from cell to cell. Principal component analysis (PCA) was performed using the RunPCA function. Cells were clustered by FindNeighbors and FindClusters functions (resolution = 0.5, dimension = 20), and visualized by t-distributed random neighbor embedding (tSNE). Marker genes for each cluster were identified by the FindAllMarkers function (logfc = 0.5, Minpct = 0.35), and cell types were annotated according to the abundance of known marker genes, as described in the previous literature [ 23 ]. The CellChat package in R was used to infer, analyze, and visualize intercellular communication in single-cell data. Machine learning Up-regulated DEGs were subjected to the least absolute shrinkage and selection operator (LASSO) regression analysis using the "glmnet" package in R. To ensure reproducibility, the number of random seeds were set to 123. A 10-fold cross-validation was implemented with logistic regression (family = "binomial") to optimize the regularization parameter (lambda.min). A support vector machine-recursive feature elimination (SVM-RFE) classification model was then performed using the "e1071" package in R, utilizing a radial basis function kernel (method = svmRadial) and 10-fold cross-validation. The "RandomForest" package in R was used to construct the Random forest (RF) model. A forest of 500 trees was trained with 10-fold cross-validation to minimize out-of-bag error, and the top 15 genes were selected based on mean decrease in Gini impurity. Diagnostic performance of signature genes was evaluated using receiver operating characteristic (ROC) curves computed via the "pROC" package in R. Area under the curve (AUC) calculation was performed on both discovery (GSE216328) and independent validation cohorts. Visualization was achieved through ggplot2. Immune infiltration analysis The immune cell-specific marker panel established by Bindea was used for immunoinfiltration analysis [ 24 ]. Single-sample gene set enrichment analysis (ssGSEA) was implemented via the GSVA R package to quantify the infiltration levels of 28 immune cell subsets within the GSE182482 dataset. The correlation patterns between the core genes and 28 immune cell subsets was analyzed by R package and visualized using the ggplot2 package. Molecular docking The 3D structures of the compounds were obtained from PubChem database ( https://pubchem.ncbi.nml.gov ). The protein structures were retrieved from Protein Data Bank database ( http://www.wwpdb.org/ ). The ligands and proteins were prepared before docking using the Prep Wiz module of the Schrödinger (v2021-3). Docking was performed using the Glide SP with all default parameters. Molecular Dynamic Simulation Molecular dynamics (MD) simulations were performed using the Desmond module in the Schrödinger suite [ 25 ]. The protein-ligand complexes were prepared using the “System Builder” function. The periodic boundary was created around the complex in a shape of a rhomboid box which was saturated with water molecules using the SPC solvation model. The system was minimized by a hybrid approach (up to 5000 iterations) of steep descent and limited-memory Broyden-Fletcher-Goldfarb-Shanno (LBFGS) algorithms. Using the OPLS4 force field, MD simulation was performed after the heating and equilibration processes. The system is normalized at equilibrium over 1000 frames in 100 ps time steps. The final production run was maintained at 100 ns with the temperature at 300 K and the pressure at 1.01325 bar for both complexes, applying the Nose-Hoover method with the NPT system. Results Construction of the herb-compound-target network QBD is composed of 13 herbs and 1 mineral. The chemical ingredient of the mineral Gypsum Fibrosum is CaSO₄·2H₂O. Considering that the specific target of a simple inorganic salt is difficult to be defined, Gypsum Fibrosum was excluded in this study. After assessing the drug-like properties of chemical ingredients, 224 molecules were obtained from 13 herbs in the QBD. Coexsiting ingredients exist in 10 herbs in the QBD (Fig. S1 ). In total, there are only 13 common ingredients present in at least 2 herbs. For examples, licorice and Scutellariae Radix share at most 4 compounds in common. The retrieved targets from TCMSP and the predicted targets by Swiss Target Prediction gave an overall of 984 potential targets for QBD. As shown in Fig. 2 , the herb-compound-target network comprises 1,156 nodes, interconnected by 7,727 edges, suggesting a possibly complicated mechanism of action of QBD in the treatment of dengue fever. Identification of DEGs Prior to bioinformatics analysis, batch effects in the validation datasets (GESE140809 and GSE182482) were evaluated, revealing that 32.1% of variability was attributed to significant batch effects. The "sva" package in R was used to correct batch effects (Fig. 3 A). Differential expression analysis was then performed on the training set GES216328 using the "DESeq2" package. A total of 1,585 DEGs, including 1,118 up-regulated and 467 down-regulated genes, were identified (Fig. 3 B). Enrichment Analysis GO enrichment analysis was performed to identify biological processes, molecular functions, and cellular components associated with DEGs. These DEGs play pivotal roles in a variety of biological processes, including cellular response to biotic stimulus, chemotaxis, protein autophosphorylation, protein tyrosine kinase activity, leukocyte apoptotic process, and regulation of smooth muscle cell proliferation (Fig. 4 A). KEGG enrichment analysis revealed that these DEGs participated in critical signal transduction pathways including cell cycle, phagocytosis of apoptotic cells, lipid and atherosclerosis, and FoxO signaling pathway (Fig. 4 B). The key KEGG pathways and related core genes were revealed by GSEA analysis. In comparison with the control group, patients with dengue fever exhibit significantly increased activity of the toll-like receptor (TLR) signaling pathway, cell cycle and RIG-I-like receptor (RLR) signaling pathway, while B cell receptor signaling and mTOR signaling pathway activities were markedly reduced (Fig. 4 C). This result indicates that QBD is able to modulate the innate immune response of host cells upon DENV infection. KEGG (B) and GSEA (C) analysis of DEGs in DENV patients. Single-cell analysis We reanalyzed the scRNA-seq data from the GEO dataset GSE271966, which contains 5,235 high-quality cells and 17,372 genes (Table S1 ). When the resolution was set to 0.5 based on the clustering tree results, these cells were divided into 10 clusters (Fig. 5 A). Figure 5 B shows the tSNE plots grouped by different experimental conditions. Cell annotation using marker genes identified 9 distinct cell types (Fig. 5 C). Bubble plots illustrate the top 5 marker genes for each cell type, with their average expression levels (Fig. 5 D). Using the CellChat package, we analyzed the complex communication networks among DENV-infected cells and identified close bidirectional interactions between 5 cell subpopulations (Fig. 5 E), despite their distinct biological functions. AUCell functional scoring of drug-targeted cells demonstrated that QBD primarily affects dendritic cells, monocytes, and macrophages (Fig. 5 F). Discovery of core targets through machine learning algorithms Since up-regulated genes and gene products are more attractive to be drug targets than down-regulated ones, we only considered up-regulated genes in the following exploration of the mechanism of action of QBD in the therapy of dengue fever. Eighty intersecting genes are derived from 1,118 up-regulated DEGs and 984 potential drug targets of QBD. Three algorithms, including LASSO, RF and SVM-RFE, were applied to identify the key functional genes. LASSO regression identified 11 potential candidate genes with diagnostic significance (Fig. 6 A), while RF and SVM-RFE detected 15 and 16 candidate genes, respectively (Fig. 6 B, 6 C). Intersection of these results from three algorithms yielded 4 shared genes: C-X-C motif chemokine 10 (CXCL10), Enhancer of Zeste Homolog 2 (EZH2), Ephrin type-B receptor 2 (EPHB2), and Low-density lipoprotein receptor (LDL) (Fig. 6 D). Differential expression analysis of 4 core genes As shown in Fig. 7 A and 7 B, CXCL10, EZH2, and LDLR exhibited significantly higher expression levels in the DENV group compared to the control group. However, LDLR expression in the validation group differed from the test group, indicating that it may not be a reliable drug target of QBD for treating dengue fever, therefore it was excluded in the subsequent studies. ROC analysis evaluated the specificity and sensitivity of the remaining three genes. The results demonstrated robust predictive performance: CXCL10 (AUC = 0.90), EZH2 (AUC = 0.84), and EPHB2 (AUC = 0.91) (Fig. 7 C, 7 D). These results confirmed the potential of these three genes as targets of QBD in treating dengue fever. Single-Gene GSEA of core genes Single-gene GSEA was performed to elucidate the roles of CXCL10, EZH2, and EPHB2 in DENV infection (Fig. 8 ). GSEA highlighted three genes with significant impacts on multiple biological processes, especially including virus-related processes such as RIG-I-like receptor signaling pathway, Epstein-Barr virus (EBV) infection, human T-cell leukemia virus 1 (HTLV-1) infection and COVID-19. Immune-related pathways, such as systemic lupus erythematosus (SLE), C-type lectin receptor (CLR) signaling, and NOD-like receptor signaling, were enriched. Additionally, pathway linked to cellular senescence was also identified. These results indicate that the core genes are closely related to DENV infection and potential targets for QBD. Immune cell infiltration analysis To explore the relationship between the core genes and immune cell infiltration, the ssGSEA algorithm was employed to investigate immune cell abundance across different groups in DENV infection. The heatmap of immune cell abundance in both control and DENV groups revealed significant increases in activated CD4 + T cells, central memory CD4 + T cells, central memory CD8 + T cells, effector memory CD4 + T cells, effector memory CD8 + T cells, γδ T cells, Th17 cells, and Th2 cells in the DENV group compared to the control group (Fig. 9 A, 9 B). Conversely, eosinophil level is markedly lower in the DENV group. These immune cells were not isolated but closely interconnected (Fig. 9 C), reflecting the immune complexity during the DENV infection. Further analysis revealed significant positive/negative correlations between the expression of core genes and these immune cells. Notably, CD4 + and CD8 + T cell subtypes exhibited strong correlations with these 3 core genes, suggesting the role QBD in immune modulation upon DENV infection (Fig. 9 D). Molecular docking To explore the core active ingredient s in QBD that contribute to the therapeutic effect against dengue fever, molecular docking was performed. QBD may play dual roles in the treatment of dengue fever, including modulating the three identified human protein targets and directly targeting the viral proteins of DENV. The three identified targets and five DENV structural and non-structural proteins were included in the molecular docking study (Fig. 10 ). Lower docking scores indicate stronger ligand-receptor interactions and higher affinity. Top 5 compounds and docking scores with their corresponding targets are listed in Table S2 . These critical compounds exhibited strong binding affinity to the targets (docking score < − 5.0 kcal/mol). It is interesting that several compounds exhibited potency to modulate two or more targets, indicating their multifaceted roles in treating dengue fever. MD simulation and interaction analysis MD simulation is a computational tool used to determine ligand binding affinity in proteins over a defined time period [ 26 ]. We performed 100 ns MD simulations for the top 5 small molecules identified from molecular docking experiments against each protein target. The root mean square deviation (RMSD) serves as a good measure of the conformational stability of the protein and ligand, indicating the extent of atomic positional deviation from the starting structure; a lower deviation signifies better conformational stability. A protein RMSD value fluctuating stably below 3 Å was considered indicative of the protein maintaining a stable conformation upon ligand binding. Representative results are shown in Fig. 10 . Eight compounds were identified to bind their corresponding target proteins stably. (+)-Catechin formed a stable complex with CXCL10 via hydrogen bonding interactions with PRO21 and VAL19 (Fig. 11 A). Dehydroglyasperins C also bind stably to CXCL10, forming hydrogen bonds with LEU65 and ILE61 (Fig. 11 B). Magnograndiolide formed a stable complex with EZH2, engaging in a hydrogen bond interaction with ASN194 (Fig. 11 C). Unfortunately, no small molecule compounds forming stable complexes with EPHB2 were identified in the MD simulations. Phillyrin formed a stable complex with the DENV E protein via hydrogen bonding interactions with ASN192 and HIS280 (Fig. 11 D). Glyasperin B formed multiple hydrogen bonding interactions with three amino acid residues (TYR137, ALA278, and HIS280) of the DENV E protein (Fig. 11 E). The DENV C protein formed a stable complex with crocetin via a hydrogen bond interaction at ALA63 (Fig. 11 F). In the interaction between the C protein and moupinamide, although no hydrogen bonding interactions were observed, the dynamics results remained stable, indicating the significant contribution of other interaction forces (Fig. 11 G). Paeoniflorin formed hydrogen bonding interactions with TYR137, ALA278, and HIS280 in the RdRp domain of DENV NS5 (Fig. 11 H). No small molecules forming stable complexes were identified for other DENV non-structural proteins, such as NS2B-NS3 and NS4B. Discussion Over the past decade, the global disease burden of dengue fever has continued to increase, with larger-scale outbreaks occurring in endemic regions. Currently, no effective drugs for the treatment of dengue fever are available. Although several drugs and therapeutic approaches have been proposed, their efficacy and/or molecular mechanisms have not been fully validated [ 27 – 29 ]. Consequently, identifying key molecular targets for dengue intervention through bioinformatics analysis holds significant clinical implications. Viral infection triggers diverse innate immune responses that can either restrict viral spread or, paradoxically, create conditions favoring virus replication. Gene enrichment analysis reveal that TLR signaling pathway and RLR signaling pathway are significantly activated upon DENV infection. Both TLR and RLR play essential roles in triggering innate immune responses upon virus infection [ 30 ]. TLR serves as primary sensors that detect evading viruses and elicit innate immune responses. They can activate NF-κB, a critical transcriptional factor, which controls the expression of a variety of inflammatory cytokine genes [ 31 ]. RIG-I are key cytoplasmic pathogen recognition receptors that are implicated in detecting viral RNAs. RIG-I binding to DENV results in the secretion of interferon by infected cells [ 32 , 33 ]. These results suggest that QBD may modulate the host immune response to DENV infection, however the detailed molecular mechanism cannot be elucidated at this stage. Highly expressed genes and gene products are more suitable than lowly expressed ones as antiviral targets for drug intervention to achieve therapeutic effects. Our study first identified 80 significantly up-regulated DEGs based on network pharmacology and bioinformatics analysis, which may be closely related to the infection mechanism of DENV. Subsequently, we employed machine learning algorithms to identify the core genes with significant discriminatory ability. By constructing machine learning models using LASSO, RF, and SVM-RFE algorithms, four shared genes (CXCL10, EZH2, EPHB2, and LDLR) were identified. Further validation ultimately revealed consistent expression patterns of CXCL10, EZH2, and EPHB2 in both the discovery and validation cohorts of DENV patients. ROC curve analysis confirmed their significant diagnostic potential in distinguishing the disease group from the control group. The AUC values demonstrated their robust predictive performance, further highlighting the crucial role of these core genes in the DENV infection process and their potential as targets of QBD. Increased CXCL10 levels are observed in patients with raised clinical severity of dengue infection, probably because CXCL10 induces excessive inflammation and increases vascular changes in the patients [ 34 ]. Proinflammatory TNFα facilitates DENV infection in endovascular dysfunction and neurotoxicity [ 35 ]. EZH2 plays an essential role in regulating the production of TNFα, a proinflammatory cytokine in lymphocytes. Huang and coauthors reported that silencing EZH2 by siEZH2 inhibited DENV2- or LPS-induced TNFα in THP-1 cells [ 36 ]. Both Ephs and ephrins have been reported to function as entry receptors for a variety of viruses, including hepatitis C virus, another member in the Flaviviridae family [ 37 , 38 ]. These previously reported experimental evidences support our bioinformatic analysis results. Molecular docking revealed that QBD achieved its therapeutic effect against dengue fever by simultaneously inhibiting both human protein targets and viral proteins. MD simulations further identified eight chemical components in QBD that can stably bind to their corresponding targets. These compounds combat dengue fever possibly through two pathways: (1) QBD inhibits the DENV-induced inflammatory responses by modulating the activity of CXCL10 and EZH2. Three compounds, including (+)-catechin, dehydroglyasperins C, and magnograndiolide, are responsible for this therapeutic pathway. (2) QBD is able to directly target DENV proteins, such as C, E and NS5, thereby inhibiting the entry and replication of DENV. Five compounds from QBD, including phillyrin, glyasperin B, crocetin, moupinamide, and paeoniflorin, contribute to this therapeutic pathway. Conclusion Our study integrated bioinformatics, machine learning and network pharmacology to identify three key targets: CXCL10, EPHB2, and EZH2 that may play essentials roles in the treatment of dengue fever by QBD. Genomic pathway analysis and immune cell profiling further clarified the crucial roles of these genes in dengue pathogenesis. Furthermore, molecular docking and MD simulations were applied to investigate the active components and mechanisms of QBD in treating dengue fever, ultimately identifying eight core active compounds. These compounds combat dengue fever through dual mechanisms: directly targeting viral proteins to inhibit entry and replication, or inhibiting the three identified key targets to mitigate DENV-induced inflammatory responses. Collectively, this work systematically delineates the molecular targets and mechanisms of QBD against dengue fever, providing a theoretical foundation for future research and clinical applications. The identified active components represent promising therapeutic candidates, highlighting their significant potential for developing novel anti-dengue therapeutics. Declarations Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Clinical trial number Not applicable Ethics approval and consent to participate Not applicable Consent for publication Not applicable Funding The authors acknowledge the financial support from the National Natural Science Foundation of China (Grant No. 22407093), ‌Basic Scientific Research Project of Liaoning Provincial Department of Education (Grant No. LJ212510163009). Author Contribution Y.X.: Conceptualization, Software, Writing - review & editing, Funding acquisition; Z.L.: Software, Formal analysis, Writing - original draft; P.Y.: Software; S.H.: Software; Y.L.: Conceptualization, Methodology, Funding acquisition, Formal analysis, Writing - review & editing; M.C.: Conceptualization, Writing - review & editing. Acknowledgements We thank the collaborating authors for their valuable contributions to this work Data Availability The datasets analyzed during the current study are available in the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE216328. References Byk LA, Gamarnik AV (2016) Properties and functions of the dengue virus capsid protein. Annu Rev Virol 3(1):263–281 Screaton G, Mongkolsapaya J, Yacoub S et al (2015) New insights into the immunopathology and control of dengue virus infection. Nat Rev Immunol 15(12):745–759 Bhatt P, Sabeena SP, Varma M et al (2021) Current understanding of the pathogenesis of dengue virus infection. Curr Microbiol 78(1):17–32 Guzman MG, Kouri G, Valdes L et al (2002) Enhanced severity of secondary dengue-2 infections: death rates in 1981 and 1997 Cuban outbreaks. Rev Panam Salud Publica 11:223–227 Sarker A, Dhama N, Gupta RD (2023) Dengue virus neutralizing antibody: a review of targets, cross-reactivity, and antibody-dependent enhancement. Front Immunol 14:1200195 Recker M, Blyuss KB, Simmons CP et al (2009) Immunological serotype interactions and their effect on the epidemiological pattern of dengue. Proc Biol Sci 276:2541–2548 Thein S, Aung MM, Shwe TN et al (1997) Risk factors in dengue shock syndrome. Am J Trop Med Hyg 56:566–572 Kok BH, Lim HT, Lim CP et al (2023) Dengue virus infection-a review of pathogenesis, vaccines, diagnosis, and therapy. Virus Res 324:199018 Troost B, Smit JM (2020) Recent advances in antiviral drug development towards dengue virus. Curr Opin Virol 43:9–21 World Health Organization (2019) Dengue vaccine: WHO position paper, September 2018-Recommendations. Vaccine 37(35):4848–4849 João EE, Lopes JR, Guedes BFR et al (2024) Advances in drug discovery of flavivirus NS2B-NS3pro serine protease inhibitors for the treatment of Dengue, Zika, and West Nile viruses. Bioorg Chem 153:107914 Pujar SAH, Sethu GV (2021) Dengue structural proteins as antiviral drug targets: Current status in the drug discovery & development. Eur J Med Chem 221:113527 Zheng Y, Liu S, Fan C et al (2020) Holistic quality evaluation of Qingwen Baidu Decoction and its anti-inflammatory effects. J Ethnopharmacol 263:113145 Xie T, Ling YK (1993) The therapeutic effect and mechanism of Qingwen Baidu Decoction on the syndrome of Qi blood double burnt caused by endotoxin in rabbits with febrile disease. Chin J Integr Tradit West Med 13(2):94–97 Wen J, Wang R, Liu H et al (2020) Potential therapeutic effect of Qingwen Baidu Decoction against Corona Virus Disease 2019: a mini review. Chin Med 15:48 Chen H, Jie C, Tang LP et al (2017) New insights into the effects and mechanism of a classic traditional Chinese medicinal formula on influenza prevention. Phytomedicine 27:52–62 Vamathevan J, Clark D, Czodrowski P et al (2019) Applications of machine learning in drug discovery and development. Nat Rev Drug Discov 18(6):463–477 Ota R, Yamashita F (2022) Application of machine learning techniques to the analysis and prediction of drug pharmacokinetics. J Control Release 352:961–969 Ru J, Li P, Wang J et al (2014) TCMSP: a database of systems pharmacology for drug discovery from herbal medicines. J Cheminform 6:13 Wang JF, Zhou H, Han LY et al (2005) TCM-ID: traditional Chinese medicine information database. Clin Pharmacol Ther 78(1):92–93 Fang S, Dong L, Liu L et al (2021) HERB: a high-throughput experiment- and reference-guided database of traditional Chinese medicine. Nucleic Acids Res 49(D1):D1197–D1206 Daina A, Michielin O, Zoete V (2017) SwissADME:a free web tool to evaluate pharmacokinetics, drug-likeness, and medicinal chemistry friendliness of small molecules. Sci Rep 7:42717 Jung M, Wells D, Rusch J et al (2019) Unified single-cell analysis of testis gene regulation and pathology in five mouse strains. eLife. ;8 Bindea G, Mlecnik B, Tosolini M et al (2013) Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity 39(4):782–795 Huang C, Li C, Choi PY et al (2011) A novel method for molecular dynamics simulation in the isothermal–isobaric ensemble. Mol Phys 109:191–202 Hou T, Wang J, Li Y et al (2011) Assessing the performance of the MM/PBSA and MM/GBSA methods. 1. The accuracy of binding free energy calculations based on molecular dynamics simulations. J Chem Inf Model 51(1):69–82 Rajapakse S, de Silva NL, Weeratunga P et al (2017) Prophylactic and therapeutic interventions for bleeding in dengue: a systematic review. Trans R Soc Trop Med Hyg 111:433–439 Suputtamongkol Y, Avirutnan P, Mairiang D et al (2021) Ivermectin accelerates circulating nonstructural protein 1 (NS1) clearance in adult dengue patients: A combined phase 2/3 randomized double-blinded placebo-controlled trial. Clin Infect Dis 72(10):e586–e593 Wilder-Smith A, Hombach J, Ferguson N et al (2019) Deliberations of the Strategic Advisory Group of Experts on Immunization on the use of CYD-TDV dengue vaccine. Lancet Infect Dis 19(1):e31–e38 Kawai T, Akira S (2008) Toll-like receptor and RIG-I-like receptor signaling. Ann N Y Acad Sci 1143:1–20 Kawai T, Akira S (2017) Signaling to NF-kappaB by Toll-like receptors. Trends Mol Med 13(11):460–469 Chazal M, Beauclair G, Gracias S et al (2018) RIG-I recognizes the 5' region of Dengue and Zika virus genomes. Cell Rep 24(2):320–328 Solstad A, Hogaboam O, Forero A et al (2022) RIG-I-like receptor regulation of immune cell function and therapeutic implications. J Immunol 209(5):845–854 Jusof FF, Lim CK, Aziz FN et al (2022) The cytokines CXCL10 and CCL2 and the kynurenine metabolite anthranilic acid accurately predict patients at risk of developing dengue with warning signs. J Infect Dis 226(11):1964–1973 Jhan MK, HuangFu WC, Chen YF et al (2018) Anti-TNF-α restricts dengue virus-induced neuropathy. J Leukoc Biol 104(5):961–968 Zhang Y, Zhang Q, Gui L et al (2018) Let-7e inhibits TNF-α expression by targeting the methyl transferase EZH2 in DENV2-infected THP-1 cells. J Cell Physiol 233(11):8605–8616 De Boer ECW, van Gils JM, van Gils MJ (2020) Ephrin-Eph signaling usage by a variety of viruses. Pharmacol Res 159:105038 Lupberger J, Zeisel MB, Xiao F et al (2011) EGFR and EphA2 are host factors for hepatitis C virus entry and possible targets for antiviral therapy. Nat Med 17(5):589–595 Additional Declarations No competing interests reported. Supplementary Files Supplementaryinformation.docx Supplementary materials Supplementary material associated with this article can be found, in the online version, at X. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 18 Mar, 2026 Reviews received at journal 17 Mar, 2026 Reviews received at journal 17 Mar, 2026 Reviewers agreed at journal 17 Mar, 2026 Reviews received at journal 16 Mar, 2026 Reviewers agreed at journal 16 Mar, 2026 Reviewers agreed at journal 16 Mar, 2026 Reviewers agreed at journal 13 Mar, 2026 Reviewers agreed at journal 11 Mar, 2026 Reviewers agreed at journal 11 Mar, 2026 Reviewers agreed at journal 11 Mar, 2026 Reviewers invited by journal 11 Mar, 2026 Editor assigned by journal 09 Jan, 2026 Submission checks completed at journal 09 Jan, 2026 First submitted to journal 05 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8520853","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":605596301,"identity":"07129e2d-2d36-47d3-a6c8-1027c0b33f71","order_by":0,"name":"Yan Xiao","email":"","orcid":"","institution":"Shenyang Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Xiao","suffix":""},{"id":605596302,"identity":"ab173603-e11d-4d0f-b69f-037db4cde06e","order_by":1,"name":"Zhanchen Liu","email":"","orcid":"","institution":"Shenyang Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Zhanchen","middleName":"","lastName":"Liu","suffix":""},{"id":605596303,"identity":"29f271bb-0f15-47ea-9f8f-65d3a6e21f6e","order_by":2,"name":"Shengjie Hu","email":"","orcid":"","institution":"Shenyang Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Shengjie","middleName":"","lastName":"Hu","suffix":""},{"id":605596305,"identity":"e91fecff-bd35-4201-9242-bcc7b014991a","order_by":3,"name":"Peng Yao","email":"","orcid":"","institution":"Shenyang Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Peng","middleName":"","lastName":"Yao","suffix":""},{"id":605596307,"identity":"3f341bc3-e958-41d4-a126-7597d4197b4d","order_by":4,"name":"Yajun Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIiWNgGAWjYBACPmYQycPAwA+kwGyGAwS0sMG0SDYQrQXGMDhAtBZ2HjOJHzKH84zPnzF7XNjGIMd3I4HxcwFeh/GYSfbwHC42u5FjbjyzjcFY8kYCs/QMAlokeHgOJ267wWMmzdvGkLjhRgJQkJAtf4BaNvefAWupJ0qLNMiWDQw5YC0JBoS1sBVby/CkJ864kVZuzHNOwnDmmYfN0vi08PMf3njzbY91Yn//4W2Pecps5PmOJx/8jE8LELBIMPaAaA4zICEBxIwN+DUAI/ADww8Qzf6MkMpRMApGwSgYoQAAkqJBOpV7LbIAAAAASUVORK5CYII=","orcid":"","institution":"Shenyang Pharmaceutical University","correspondingAuthor":true,"prefix":"","firstName":"Yajun","middleName":"","lastName":"Liu","suffix":""},{"id":605596309,"identity":"e7d3bd30-214f-42a6-a800-58fc7314ba91","order_by":5,"name":"Maosheng Cheng","email":"","orcid":"","institution":"Shenyang Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Maosheng","middleName":"","lastName":"Cheng","suffix":""}],"badges":[],"createdAt":"2026-01-05 11:38:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8520853/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8520853/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104808535,"identity":"1e08f072-a263-4d83-a86a-75f6127387cb","added_by":"auto","created_at":"2026-03-17 12:38:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3378573,"visible":true,"origin":"","legend":"\u003cp\u003eOverall workflow of the research design.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/8358cafdd52968e61cc72d68.png"},{"id":104699726,"identity":"2e689886-3029-40b9-a8d1-12e2958096af","added_by":"auto","created_at":"2026-03-16 08:22:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":9572419,"visible":true,"origin":"","legend":"\u003cp\u003eHerb-­compound-­target network.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/e1d28f45e2e79c25f1e6f587.png"},{"id":104699718,"identity":"a197a072-68ea-457c-b067-631a00367746","added_by":"auto","created_at":"2026-03-16 08:22:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":175543,"visible":true,"origin":"","legend":"\u003cp\u003eDEG analysis of potential targets of QBD. (A) Batch effect removal and validation of DEGs. PCA plots depict the expression patterns within two DENV datasets before and after the removal of batch effects. (B) Volcano plot illustrating DEGs in the DENV group.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/282678004f0a65fc04d53e06.png"},{"id":104699723,"identity":"6c67ab5d-728d-4c2e-bc45-4940c8ec98a3","added_by":"auto","created_at":"2026-03-16 08:22:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":493426,"visible":true,"origin":"","legend":"\u003cp\u003eInvestigation of pathological pathways involved in dengue fever patients. GO (A),\u003c/p\u003e\n\u003cp\u003eKEGG (B) and GSEA (C) analysis of DEGs in DENV patients.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/3d116c4e25c90a96fb1ed5b1.png"},{"id":104782610,"identity":"180a61c5-5598-4ebe-a08f-080e110aa5e8","added_by":"auto","created_at":"2026-03-17 07:57:36","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":603423,"visible":true,"origin":"","legend":"\u003cp\u003eThe analysis of cell types that QBD may influence based on scRNA-seq data. (A) The tSNE plot of all cells colored by cluster. (B) A tSNE plot showing the source of the original sample (C) Annotation map of cell types. (D) The top 5 marker genes in each cell type. (E) Intercellular communication networks. (F) Pathways of action of QBD.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/52f3365d34d2364a53383435.png"},{"id":104783137,"identity":"25ad1c89-2463-452c-970f-03c6577f79a8","added_by":"auto","created_at":"2026-03-17 07:58:16","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":296889,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification of core genes for QBD against DENV through machine learning algorithms.\u003cstrong\u003e \u003c/strong\u003eGenes identified by the LASSO logistic regression algorithm (A), SVM-RFE algorithm (B) and RF algorithm (C). Venn diagram illustrates the intersection of genes from three algorithms (D).\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/d3ba20b0992d8f53b48c9a00.png"},{"id":104699720,"identity":"bc4980df-c816-4bfd-b23a-9306a544fab1","added_by":"auto","created_at":"2026-03-16 08:22:13","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1188663,"visible":true,"origin":"","legend":"\u003cp\u003eDifferential expression analysis of CXCL10, EZH2 and EPHB2. (A) Training group. (B) Validation group. (C) ROC curve of the training group. (D) ROC curve of the validation group.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/c56188349d1cf7517bc0e284.png"},{"id":104783138,"identity":"c3d98442-9fc1-49a5-b69d-4ed05ee0cddd","added_by":"auto","created_at":"2026-03-17 07:58:16","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":1154980,"visible":true,"origin":"","legend":"\u003cp\u003eGSEA enrichment pathways for core genes in DENV infection. (A) CXCL10, (B) EZH2, (C) EPHB2.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/f59c3dd7df1e62486f2c3ae8.png"},{"id":104699725,"identity":"f64bb651-a79d-4b55-8688-6e7f73ca3d48","added_by":"auto","created_at":"2026-03-16 08:22:13","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":2416541,"visible":true,"origin":"","legend":"\u003cp\u003eDENV immune cell infiltration analysis.\u003cstrong\u003e \u003c/strong\u003e(A) Percentage of various immune cells in the immune microenvironment of Control and DNEV. (B) Violin diagram indicating the comparison of 28 types of immune cells between the DENV and control groups. (C) Correlations between immune cells. (D) Correlation between both core gene expression and immune cells in the DENV group.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/6de0cfb49917ad6829514412.png"},{"id":104699727,"identity":"bb71a4af-e213-424d-894b-14e6a2ebb351","added_by":"auto","created_at":"2026-03-16 08:22:13","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":1651813,"visible":true,"origin":"","legend":"\u003cp\u003eThe heatmap for the docking scores (kcal/mol) of chemical components in QBD towards three hub targets and DENV proteins.\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/d24a67b4e5daee55574e6863.png"},{"id":104699728,"identity":"f40cb3fc-074d-42cf-b26a-4005558ebf56","added_by":"auto","created_at":"2026-03-16 08:22:14","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":6053829,"visible":true,"origin":"","legend":"\u003cp\u003eThe representative result for the binding mode of chemical components in QBD with proteins, and RMSD fluctuation of the complexes in MD simulation. (A) CXCL10- (+)-catechin complex. (B) CXCL10-dehydroglyasperins C complex. (C) EZH2-magnograndiolide complex. (D) DENV envelope protein-phillyrin complex. (E) DENV envelope protein-glyasperin B complex. (F) DENV capsid protein-crocetin complex. (G) DENV capsid protein-moupinamide complex. (H) DENV NS5- paeoniflorin complex.\u003c/p\u003e","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/8cdd43da246c18f12e37c8c1.png"},{"id":104809784,"identity":"8547a629-4d83-4d7f-a0c5-53887d735020","added_by":"auto","created_at":"2026-03-17 12:53:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":27138964,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/7b7dfc93-32f4-4b09-82ec-1a32c020bb21.pdf"},{"id":104783156,"identity":"80481ef0-7cee-42b1-a73f-e113f29601a7","added_by":"auto","created_at":"2026-03-17 07:58:18","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":93118,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary materials\u003c/p\u003e\n\u003cp\u003eSupplementary material associated with this article can be found, in the online version, at X.\u003c/p\u003e","description":"","filename":"Supplementaryinformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-8520853/v1/0b8218bdc99d1438db2e8695.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Unveiling the molecular mechanism of Qing Wen Bai Du decoction against dengue fever: An integrated study of bioinformatic analysis, machine learning and network pharmacology","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDengue fever is an acute mosquito-borne infectious disease caused by the dengue virus (DENV). It is primarily transmitted through the bites of \u003cem\u003eAedes aegypti\u003c/em\u003e and \u003cem\u003eAedes albopictus\u003c/em\u003e mosquitoes. The main epidemic areas are tropical and subtropical regions, such as Southeast Asia, the Caribbean, South America and South China. The world health organization (WHO) estimates that about half of the world's population is now at risk of dengue fever with 100\u0026ndash;400\u0026nbsp;million infections occurring each year. The highest number of dengue cases was recorded in 2023, affecting over 80 countries and resulting in over 6.5\u0026nbsp;million cases and more than 7300 dengue-related deaths.\u003c/p\u003e \u003cp\u003eDENV may cause a primary infection or a secondary infection, the latter of which is called as dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS). DHF proves life-threatening due to a wide range of clinical manifestations, including plasma leakage, fluid accumulation, respiratory distress, hemorrhage, and organ damage. Without proper treatment, the mortality rate of DHF may exceed 20%.\u003c/p\u003e \u003cp\u003eDENV is a positive-sense single-stranded RNA virus belonging to the \u003cem\u003eFlaviviridae\u003c/em\u003e family [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. DENV genome encodes for three structural proteins (C, prM, and E) and seven non-structural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Based on structural differences on their surface, DENV is classified into four distinct serotypes (DENV-1, DENV-2, DENV-3, and DENV-4). Each serotype can cause a spectrum of disease manifestations. Infection with one serotype confers protective immunity against that specific serotype but does not provide cross-protective immunity against other serotypes. Notably, studies have shown that secondary infections with heterologous serotypes are a significant risk factor for developing DHF/DSS [\u003cspan additionalcitationids=\"CR5 CR6\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOver the past decades, a great effort has been devoted into the research of vaccines and anti-viral drugs against DENV, however there are currently no approved antivirals available to treat dengue fever [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Dengvaxia is the world's first vaccine for DENV, however it is only approved for use in individuals 6 through 16 years of age with laboratory-confirmed previous dengue infection and living in endemic areas [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Generally, antiviral drug discovery can be approached in two ways, including inhibitors that target viral components and inhibitors that target host cell factors based on the inhibitory mechanism. Some drugs targeting non-structural proteins, such as temoporfin, niclosamide, and nitazoxanide, are currently undergoing clinical trials for the treatment of dengue fever [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAccording to the fundamental theory of traditional Chinese medicine, dengue fever belongs to the category of \"plague\" and \"epidemic\", which has the characteristics of \"dampness, toxicity, stasis and closure\". As mentioned in \"The Diagnosis and Treatment Guidelines for Dengue Fever\" issued by the National Health Commission of China, Qingwen Baidu Decoction (QBD) is used to treat the critical phase of dengue fever [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Originating from the Qing dynasty, QBD is renowned for its effects in heat-clearing and detoxifying, making it one of the most representative antiviral formulas in traditional Chinese medicine [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. QBD is composed of 14 traditional Chinese medicinal ingredients: Gypsum Fibrosum, Rhinoceros Unicornis, Rehmannia Glutinosa, Gardeniae Fructus, Coptidis Rhizoma, Platycodon Grandiforus, Anemarrhenae Rhizoma, Scutellariae Radix, Radix Paeoniae Rubra, Moutan Cortex, Figwort Root, Phyllostachys Nigra, Forsythiae Fructus and licorice [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eModern pharmacological studies indicate that QBD is capable of blocking viral entry into cells, inhibiting viral replication, and demonstrating significant suppressive effects against multiple influenza viruses. Moreover, QBD also improves coagulation dysfunction, mitigates organ damage, alleviates clinical symptoms, and enhances patient prognosis [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. However, the specific active chemical components responsible for its efficacy against dengue fever, as well as the molecular mechanisms underlying its antiviral effects, remain unclear.\u003c/p\u003e \u003cp\u003eMachine learning can provide decision-making support for specific challenges in drug research by analyzing large amounts of bioinformatics data. It has been applied across various stages of drug discovery, including target identification and validation, recognition of prognostic biomarkers, and drug design and screening [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Machine learning aids in identifying potential disease-related molecules and signaling pathways, pinpointing critical therapeutic targets, thereby offering vital insights for novel drug discovery and elucidating the mechanisms of existing drugs [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOur work described here integrates bioinformatics analysis, machine learning and network pharmacology to identify potential therapeutic targets and associated signaling pathways for adjuvant therapeutic pathways of QBD against dengue fever. Subsequently, these potential targets are combined with dengue virus proteins to identify the active chemical ingredients and explain the molecular mechanisms of QBD. The overall workflow of the research design is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eHerb-compound-target network construction\u003c/p\u003e \u003cp\u003eNatural chemical compounds in the QBD were retrieved from the databases including TCMSP, TCMID and Herb [\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. These compounds were initially screened by the following criteria: oral bioavailability (OB\u0026thinsp;\u0026gt;\u0026thinsp;30%) and drug-likeness (DL\u0026thinsp;\u0026gt;\u0026thinsp;0.18). Subsequently, SwissADME [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] was employed to filter compounds adhering to the \"rule of five\" principles: molecular weight\u0026thinsp;\u0026lt;\u0026thinsp;500 Da, hydrogen bond donors (HBDs)\u0026thinsp;\u0026le;\u0026thinsp;5, hydrogen bond acceptors (HBAs)\u0026thinsp;\u0026le;\u0026thinsp;10, lipophilicity (logP)\u0026thinsp;\u0026lt;\u0026thinsp;5, and rotatable bonds\u0026thinsp;\u0026le;\u0026thinsp;10. Potential targets of the obtained compounds were collected from the databases and predicted by the SwissTargetPrediction. The combined targets were used to construct the herb-compound-target network by Cytoscape (version 3.10.1), visualizing interactions between QBD and predicted biological targets.\u003c/p\u003e \u003cp\u003eDifferential gene expression analysis\u003c/p\u003e \u003cp\u003eThe DENV-3 dataset (GSE216328) was retrieved from the Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). After data preprocessing, including normalization and batch effect correction, differential expression analysis was performed using the R software package DESeq2. Differentially expressed genes (DEGs) were defined using the following criteria: adjusted \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and |log2 fold change (FC)| \u0026gt; 0.5. Volcano plots were generated to visualize the distribution of DEGs based on significance and fold change values, highlighting upregulated (red) and downregulated (blue) genes.\u003c/p\u003e \u003cp\u003eGene Set Enrichment Analysis\u003c/p\u003e \u003cp\u003eGene Set Enrichment Analysis (GSEA) was performed using GSEA v4.3.2 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/gsea/index.jsp\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/gsea/index.jsp\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to evaluate the distribution patterns of DEGs within predefined gene sets, ranked by their phenotypic relevance, to identify their contributions in DENV infection. Gene expression levels were sorted in descending order. The c2.cp.v7.2.symbols.gmt gene set collection was used from the Molecular Signatures Database (MSigDB) to assess enrichment significance. Statistical thresholds were set as false discovery rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.25. Enriched pathways were visualized using GSEA-generated enrichment plots and heatmaps. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted on the DEGs using R packages \"ggplot2 \" and \"pathview\", with significance defined as adjusted P\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003cp\u003escRNA-seq analysis\u003c/p\u003e \u003cp\u003eDengue virus single-cell transcriptome data was obtained from the GSE271966 dataset. The single-cell RNA sequencing (scRNA-seq) data was analyzed using the Seurat package in R. Genes expressed in \u0026lt;\u0026thinsp;3 cells and cells detectable in \u0026lt;\u0026thinsp;300 genes were removed. Data processing was performed using a single-cell matrix for each sample, and log-normalization was performed to standardize the data for each sample. The FindVariableFeatures function was used to obtain features that vary significantly from cell to cell. Principal component analysis (PCA) was performed using the RunPCA function. Cells were clustered by FindNeighbors and FindClusters functions (resolution\u0026thinsp;=\u0026thinsp;0.5, dimension\u0026thinsp;=\u0026thinsp;20), and visualized by t-distributed random neighbor embedding (tSNE). Marker genes for each cluster were identified by the FindAllMarkers function (logfc\u0026thinsp;=\u0026thinsp;0.5, Minpct\u0026thinsp;=\u0026thinsp;0.35), and cell types were annotated according to the abundance of known marker genes, as described in the previous literature [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. The CellChat package in R was used to infer, analyze, and visualize intercellular communication in single-cell data.\u003c/p\u003e \u003cp\u003eMachine learning\u003c/p\u003e \u003cp\u003eUp-regulated DEGs were subjected to the least absolute shrinkage and selection operator (LASSO) regression analysis using the \"glmnet\" package in R. To ensure reproducibility, the number of random seeds were set to 123. A 10-fold cross-validation was implemented with logistic regression (family = \"binomial\") to optimize the regularization parameter (lambda.min). A support vector machine-recursive feature elimination (SVM-RFE) classification model was then performed using the \"e1071\" package in R, utilizing a radial basis function kernel (method\u0026thinsp;=\u0026thinsp;svmRadial) and 10-fold cross-validation. The \"RandomForest\" package in R was used to construct the Random forest (RF) model. A forest of 500 trees was trained with 10-fold cross-validation to minimize out-of-bag error, and the top 15 genes were selected based on mean decrease in Gini impurity. Diagnostic performance of signature genes was evaluated using receiver operating characteristic (ROC) curves computed via the \"pROC\" package in R. Area under the curve (AUC) calculation was performed on both discovery (GSE216328) and independent validation cohorts. Visualization was achieved through ggplot2.\u003c/p\u003e \u003cp\u003eImmune infiltration analysis\u003c/p\u003e \u003cp\u003eThe immune cell-specific marker panel established by Bindea was used for immunoinfiltration analysis [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Single-sample gene set enrichment analysis (ssGSEA) was implemented via the GSVA R package to quantify the infiltration levels of 28 immune cell subsets within the GSE182482 dataset. The correlation patterns between the core genes and 28 immune cell subsets was analyzed by R package and visualized using the ggplot2 package.\u003c/p\u003e \u003cp\u003eMolecular docking\u003c/p\u003e \u003cp\u003eThe 3D structures of the compounds were obtained from PubChem database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubchem.ncbi.nml.gov\u003c/span\u003e\u003cspan address=\"https://pubchem.ncbi.nml.gov\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The protein structures were retrieved from Protein Data Bank database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.wwpdb.org/\u003c/span\u003e\u003cspan address=\"http://www.wwpdb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The ligands and proteins were prepared before docking using the Prep Wiz module of the Schr\u0026ouml;dinger (v2021-3). Docking was performed using the Glide SP with all default parameters.\u003c/p\u003e \u003cp\u003eMolecular Dynamic Simulation\u003c/p\u003e \u003cp\u003eMolecular dynamics (MD) simulations were performed using the Desmond module in the Schr\u0026ouml;dinger suite [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. The protein-ligand complexes were prepared using the \u0026ldquo;System Builder\u0026rdquo; function. The periodic boundary was created around the complex in a shape of a rhomboid box which was saturated with water molecules using the SPC solvation model. The system was minimized by a hybrid approach (up to 5000 iterations) of steep descent and limited-memory Broyden-Fletcher-Goldfarb-Shanno (LBFGS) algorithms. Using the OPLS4 force field, MD simulation was performed after the heating and equilibration processes. The system is normalized at equilibrium over 1000 frames in 100 ps time steps. The final production run was maintained at 100 ns with the temperature at 300 K and the pressure at 1.01325 bar for both complexes, applying the Nose-Hoover method with the NPT system.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eConstruction of the herb-compound-target network\u003c/p\u003e \u003cp\u003eQBD is composed of 13 herbs and 1 mineral. The chemical ingredient of the mineral Gypsum Fibrosum is CaSO₄\u0026middot;2H₂O. Considering that the specific target of a simple inorganic salt is difficult to be defined, Gypsum Fibrosum was excluded in this study. After assessing the drug-like properties of chemical ingredients, 224 molecules were obtained from 13 herbs in the QBD. Coexsiting ingredients exist in 10 herbs in the QBD (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). In total, there are only 13 common ingredients present in at least 2 herbs. For examples, licorice and Scutellariae Radix share at most 4 compounds in common. The retrieved targets from TCMSP and the predicted targets by Swiss Target Prediction gave an overall of 984 potential targets for QBD. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, the herb-compound-target network comprises 1,156 nodes, interconnected by 7,727 edges, suggesting a possibly complicated mechanism of action of QBD in the treatment of dengue fever.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIdentification of DEGs\u003c/p\u003e \u003cp\u003ePrior to bioinformatics analysis, batch effects in the validation datasets (GESE140809 and GSE182482) were evaluated, revealing that 32.1% of variability was attributed to significant batch effects. The \"sva\" package in R was used to correct batch effects (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Differential expression analysis was then performed on the training set GES216328 using the \"DESeq2\" package. A total of 1,585 DEGs, including 1,118 up-regulated and 467 down-regulated genes, were identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eEnrichment Analysis\u003c/p\u003e \u003cp\u003eGO enrichment analysis was performed to identify biological processes, molecular functions, and cellular components associated with DEGs. These DEGs play pivotal roles in a variety of biological processes, including cellular response to biotic stimulus, chemotaxis, protein autophosphorylation, protein tyrosine kinase activity, leukocyte apoptotic process, and regulation of smooth muscle cell proliferation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). KEGG enrichment analysis revealed that these DEGs participated in critical signal transduction pathways including cell cycle, phagocytosis of apoptotic cells, lipid and atherosclerosis, and FoxO signaling pathway (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). The key KEGG pathways and related core genes were revealed by GSEA analysis. In comparison with the control group, patients with dengue fever exhibit significantly increased activity of the toll-like receptor (TLR) signaling pathway, cell cycle and RIG-I-like receptor (RLR) signaling pathway, while B cell receptor signaling and mTOR signaling pathway activities were markedly reduced (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). This result indicates that QBD is able to modulate the innate immune response of host cells upon DENV infection.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eKEGG (B) and GSEA (C) analysis of DEGs in DENV patients.\u003c/p\u003e \u003cp\u003eSingle-cell analysis\u003c/p\u003e \u003cp\u003eWe reanalyzed the scRNA-seq data from the GEO dataset GSE271966, which contains 5,235 high-quality cells and 17,372 genes (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). When the resolution was set to 0.5 based on the clustering tree results, these cells were divided into 10 clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB shows the tSNE plots grouped by different experimental conditions. Cell annotation using marker genes identified 9 distinct cell types (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). Bubble plots illustrate the top 5 marker genes for each cell type, with their average expression levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). Using the CellChat package, we analyzed the complex communication networks among DENV-infected cells and identified close bidirectional interactions between 5 cell subpopulations (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE), despite their distinct biological functions. AUCell functional scoring of drug-targeted cells demonstrated that QBD primarily affects dendritic cells, monocytes, and macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDiscovery of core targets through machine learning algorithms\u003c/p\u003e \u003cp\u003eSince up-regulated genes and gene products are more attractive to be drug targets than down-regulated ones, we only considered up-regulated genes in the following exploration of the mechanism of action of QBD in the therapy of dengue fever. Eighty intersecting genes are derived from 1,118 up-regulated DEGs and 984 potential drug targets of QBD. Three algorithms, including LASSO, RF and SVM-RFE, were applied to identify the key functional genes. LASSO regression identified 11 potential candidate genes with diagnostic significance (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA), while RF and SVM-RFE detected 15 and 16 candidate genes, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC). Intersection of these results from three algorithms yielded 4 shared genes: C-X-C motif chemokine 10 (CXCL10), Enhancer of Zeste Homolog 2 (EZH2), Ephrin type-B receptor 2 (EPHB2), and Low-density lipoprotein receptor (LDL) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDifferential expression analysis of 4 core genes\u003c/p\u003e \u003cp\u003eAs shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA and \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB, CXCL10, EZH2, and LDLR exhibited significantly higher expression levels in the DENV group compared to the control group. However, LDLR expression in the validation group differed from the test group, indicating that it may not be a reliable drug target of QBD for treating dengue fever, therefore it was excluded in the subsequent studies. ROC analysis evaluated the specificity and sensitivity of the remaining three genes. The results demonstrated robust predictive performance: CXCL10 (AUC\u0026thinsp;=\u0026thinsp;0.90), EZH2 (AUC\u0026thinsp;=\u0026thinsp;0.84), and EPHB2 (AUC\u0026thinsp;=\u0026thinsp;0.91) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC, \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD). These results confirmed the potential of these three genes as targets of QBD in treating dengue fever.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSingle-Gene GSEA of core genes\u003c/p\u003e \u003cp\u003eSingle-gene GSEA was performed to elucidate the roles of CXCL10, EZH2, and EPHB2 in DENV infection (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). GSEA highlighted three genes with significant impacts on multiple biological processes, especially including virus-related processes such as RIG-I-like receptor signaling pathway, Epstein-Barr virus (EBV) infection, human T-cell leukemia virus 1 (HTLV-1) infection and COVID-19. Immune-related pathways, such as systemic lupus erythematosus (SLE), C-type lectin receptor (CLR) signaling, and NOD-like receptor signaling, were enriched. Additionally, pathway linked to cellular senescence was also identified. These results indicate that the core genes are closely related to DENV infection and potential targets for QBD.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eImmune cell infiltration analysis\u003c/p\u003e \u003cp\u003eTo explore the relationship between the core genes and immune cell infiltration, the ssGSEA algorithm was employed to investigate immune cell abundance across different groups in DENV infection. The heatmap of immune cell abundance in both control and DENV groups revealed significant increases in activated CD4\u003csup\u003e+\u003c/sup\u003e T cells, central memory CD4\u003csup\u003e+\u003c/sup\u003e T cells, central memory CD8\u0026thinsp;+\u0026thinsp;T cells, effector memory CD4\u003csup\u003e+\u003c/sup\u003e T cells, effector memory CD8\u003csup\u003e+\u003c/sup\u003e T cells, γδ T cells, Th17 cells, and Th2 cells in the DENV group compared to the control group (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB). Conversely, eosinophil level is markedly lower in the DENV group. These immune cells were not isolated but closely interconnected (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eC), reflecting the immune complexity during the DENV infection. Further analysis revealed significant positive/negative correlations between the expression of core genes and these immune cells. Notably, CD4\u003csup\u003e+\u003c/sup\u003e and CD8\u003csup\u003e+\u003c/sup\u003e T cell subtypes exhibited strong correlations with these 3 core genes, suggesting the role QBD in immune modulation upon DENV infection (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMolecular docking\u003c/p\u003e \u003cp\u003eTo explore the core active ingredient\u003cb\u003es\u003c/b\u003e in QBD that contribute to the therapeutic effect against dengue fever, molecular docking was performed. QBD may play dual roles in the treatment of dengue fever, including modulating the three identified human protein targets and directly targeting the viral proteins of DENV. The three identified targets and five DENV structural and non-structural proteins were included in the molecular docking study (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e). Lower docking scores indicate stronger ligand-receptor interactions and higher affinity. Top 5 compounds and docking scores with their corresponding targets are listed in Table \u003cb\u003eS2\u003c/b\u003e. These critical compounds exhibited strong binding affinity to the targets (docking score\u0026thinsp;\u0026lt;\u0026thinsp;\u0026minus;\u0026thinsp;5.0 kcal/mol). It is interesting that several compounds exhibited potency to modulate two or more targets, indicating their multifaceted roles in treating dengue fever.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMD simulation and interaction analysis\u003c/p\u003e \u003cp\u003eMD simulation is a computational tool used to determine ligand binding affinity in proteins over a defined time period [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. We performed 100 ns MD simulations for the top 5 small molecules identified from molecular docking experiments against each protein target. The root mean square deviation (RMSD) serves as a good measure of the conformational stability of the protein and ligand, indicating the extent of atomic positional deviation from the starting structure; a lower deviation signifies better conformational stability. A protein RMSD value fluctuating stably below 3 \u0026Aring; was considered indicative of the protein maintaining a stable conformation upon ligand binding. Representative results are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e. Eight compounds were identified to bind their corresponding target proteins stably. (+)-Catechin formed a stable complex with CXCL10 via hydrogen bonding interactions with PRO21 and VAL19 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eA). Dehydroglyasperins C also bind stably to CXCL10, forming hydrogen bonds with LEU65 and ILE61 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eB). Magnograndiolide formed a stable complex with EZH2, engaging in a hydrogen bond interaction with ASN194 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eC). Unfortunately, no small molecule compounds forming stable complexes with EPHB2 were identified in the MD simulations. Phillyrin formed a stable complex with the DENV E protein via hydrogen bonding interactions with ASN192 and HIS280 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eD). Glyasperin B formed multiple hydrogen bonding interactions with three amino acid residues (TYR137, ALA278, and HIS280) of the DENV E protein (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eE). The DENV C protein formed a stable complex with crocetin via a hydrogen bond interaction at ALA63 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eF). In the interaction between the C protein and moupinamide, although no hydrogen bonding interactions were observed, the dynamics results remained stable, indicating the significant contribution of other interaction forces (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eG). Paeoniflorin formed hydrogen bonding interactions with TYR137, ALA278, and HIS280 in the RdRp domain of DENV NS5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eH). No small molecules forming stable complexes were identified for other DENV non-structural proteins, such as NS2B-NS3 and NS4B.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOver the past decade, the global disease burden of dengue fever has continued to increase, with larger-scale outbreaks occurring in endemic regions. Currently, no effective drugs for the treatment of dengue fever are available. Although several drugs and therapeutic approaches have been proposed, their efficacy and/or molecular mechanisms have not been fully validated [\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Consequently, identifying key molecular targets for dengue intervention through bioinformatics analysis holds significant clinical implications.\u003c/p\u003e \u003cp\u003eViral infection triggers diverse innate immune responses that can either restrict viral spread or, paradoxically, create conditions favoring virus replication. Gene enrichment analysis reveal that TLR signaling pathway and RLR signaling pathway are significantly activated upon DENV infection. Both TLR and RLR play essential roles in triggering innate immune responses upon virus infection [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. TLR serves as primary sensors that detect evading viruses and elicit innate immune responses. They can activate NF-κB, a critical transcriptional factor, which controls the expression of a variety of inflammatory cytokine genes [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. RIG-I are key cytoplasmic pathogen recognition receptors that are implicated in detecting viral RNAs. RIG-I binding to DENV results in the secretion of interferon by infected cells [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. These results suggest that QBD may modulate the host immune response to DENV infection, however the detailed molecular mechanism cannot be elucidated at this stage.\u003c/p\u003e \u003cp\u003eHighly expressed genes and gene products are more suitable than lowly expressed ones as antiviral targets for drug intervention to achieve therapeutic effects. Our study first identified 80 significantly up-regulated DEGs based on network pharmacology and bioinformatics analysis, which may be closely related to the infection mechanism of DENV. Subsequently, we employed machine learning algorithms to identify the core genes with significant discriminatory ability. By constructing machine learning models using LASSO, RF, and SVM-RFE algorithms, four shared genes (CXCL10, EZH2, EPHB2, and LDLR) were identified. Further validation ultimately revealed consistent expression patterns of CXCL10, EZH2, and EPHB2 in both the discovery and validation cohorts of DENV patients. ROC curve analysis confirmed their significant diagnostic potential in distinguishing the disease group from the control group. The AUC values demonstrated their robust predictive performance, further highlighting the crucial role of these core genes in the DENV infection process and their potential as targets of QBD.\u003c/p\u003e \u003cp\u003eIncreased CXCL10 levels are observed in patients with raised clinical severity of dengue infection, probably because CXCL10 induces excessive inflammation and increases vascular changes in the patients [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Proinflammatory TNFα facilitates DENV infection in endovascular dysfunction and neurotoxicity [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. EZH2 plays an essential role in regulating the production of TNFα, a proinflammatory cytokine in lymphocytes. Huang and coauthors reported that silencing EZH2 by siEZH2 inhibited DENV2- or LPS-induced TNFα in THP-1 cells [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Both Ephs and ephrins have been reported to function as entry receptors for a variety of viruses, including hepatitis C virus, another member in the Flaviviridae family [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. These previously reported experimental evidences support our bioinformatic analysis results.\u003c/p\u003e \u003cp\u003eMolecular docking revealed that QBD achieved its therapeutic effect against dengue fever by simultaneously inhibiting both human protein targets and viral proteins. MD simulations further identified eight chemical components in QBD that can stably bind to their corresponding targets. These compounds combat dengue fever possibly through two pathways: (1) QBD inhibits the DENV-induced inflammatory responses by modulating the activity of CXCL10 and EZH2. Three compounds, including (+)-catechin, dehydroglyasperins C, and magnograndiolide, are responsible for this therapeutic pathway. (2) QBD is able to directly target DENV proteins, such as C, E and NS5, thereby inhibiting the entry and replication of DENV. Five compounds from QBD, including phillyrin, glyasperin B, crocetin, moupinamide, and paeoniflorin, contribute to this therapeutic pathway.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur study integrated bioinformatics, machine learning and network pharmacology to identify three key targets: CXCL10, EPHB2, and EZH2 that may play essentials roles in the treatment of dengue fever by QBD. Genomic pathway analysis and immune cell profiling further clarified the crucial roles of these genes in dengue pathogenesis. Furthermore, molecular docking and MD simulations were applied to investigate the active components and mechanisms of QBD in treating dengue fever, ultimately identifying eight core active compounds. These compounds combat dengue fever through dual mechanisms: directly targeting viral proteins to inhibit entry and replication, or inhibiting the three identified key targets to mitigate DENV-induced inflammatory responses. Collectively, this work systematically delineates the molecular targets and mechanisms of QBD against dengue fever, providing a theoretical foundation for future research and clinical applications. The identified active components represent promising therapeutic candidates, highlighting their significant potential for developing novel anti-dengue therapeutics.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eDeclaration of competing interest\u003c/h2\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003ch2\u003eClinical trial number\u003c/h2\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eThe authors acknowledge the financial support from the National Natural Science Foundation of China (Grant No. 22407093), \u0026zwnj;Basic Scientific Research Project of Liaoning Provincial Department of Education (Grant No. LJ212510163009).\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eY.X.: Conceptualization, Software, Writing - review \u0026amp; editing, Funding acquisition; Z.L.: Software, Formal analysis, Writing - original draft; P.Y.: Software; S.H.: Software; Y.L.: Conceptualization, Methodology, Funding acquisition, Formal analysis, Writing - review \u0026amp; editing; M.C.: Conceptualization, Writing - review \u0026amp; editing.\u003c/p\u003e\n\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eWe thank the collaborating authors for their valuable contributions to this work\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe datasets analyzed during the current study are available in the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE216328.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eByk LA, Gamarnik AV (2016) Properties and functions of the dengue virus capsid protein. Annu Rev Virol 3(1):263\u0026ndash;281\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScreaton G, Mongkolsapaya J, Yacoub S et al (2015) New insights into the immunopathology and control of dengue virus infection. Nat Rev Immunol 15(12):745\u0026ndash;759\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhatt P, Sabeena SP, Varma M et al (2021) Current understanding of the pathogenesis of dengue virus infection. Curr Microbiol 78(1):17\u0026ndash;32\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuzman MG, Kouri G, Valdes L et al (2002) Enhanced severity of secondary dengue-2 infections: death rates in 1981 and 1997 Cuban outbreaks. Rev Panam Salud Publica 11:223\u0026ndash;227\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSarker A, Dhama N, Gupta RD (2023) Dengue virus neutralizing antibody: a review of targets, cross-reactivity, and antibody-dependent enhancement. Front Immunol 14:1200195\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRecker M, Blyuss KB, Simmons CP et al (2009) Immunological serotype interactions and their effect on the epidemiological pattern of dengue. Proc Biol Sci 276:2541\u0026ndash;2548\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThein S, Aung MM, Shwe TN et al (1997) Risk factors in dengue shock syndrome. Am J Trop Med Hyg 56:566\u0026ndash;572\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKok BH, Lim HT, Lim CP et al (2023) Dengue virus infection-a review of pathogenesis, vaccines, diagnosis, and therapy. Virus Res 324:199018\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTroost B, Smit JM (2020) Recent advances in antiviral drug development towards dengue virus. Curr Opin Virol 43:9\u0026ndash;21\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Health Organization (2019) Dengue vaccine: WHO position paper, September 2018-Recommendations. Vaccine 37(35):4848\u0026ndash;4849\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJo\u0026atilde;o EE, Lopes JR, Guedes BFR et al (2024) Advances in drug discovery of flavivirus NS2B-NS3pro serine protease inhibitors for the treatment of Dengue, Zika, and West Nile viruses. Bioorg Chem 153:107914\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePujar SAH, Sethu GV (2021) Dengue structural proteins as antiviral drug targets: Current status in the drug discovery \u0026amp; development. Eur J Med Chem 221:113527\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng Y, Liu S, Fan C et al (2020) Holistic quality evaluation of Qingwen Baidu Decoction and its anti-inflammatory effects. J Ethnopharmacol 263:113145\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXie T, Ling YK (1993) The therapeutic effect and mechanism of Qingwen Baidu Decoction on the syndrome of Qi blood double burnt caused by endotoxin in rabbits with febrile disease. Chin J Integr Tradit West Med 13(2):94\u0026ndash;97\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWen J, Wang R, Liu H et al (2020) Potential therapeutic effect of Qingwen Baidu Decoction against Corona Virus Disease 2019: a mini review. Chin Med 15:48\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen H, Jie C, Tang LP et al (2017) New insights into the effects and mechanism of a classic traditional Chinese medicinal formula on influenza prevention. Phytomedicine 27:52\u0026ndash;62\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVamathevan J, Clark D, Czodrowski P et al (2019) Applications of machine learning in drug discovery and development. Nat Rev Drug Discov 18(6):463\u0026ndash;477\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOta R, Yamashita F (2022) Application of machine learning techniques to the analysis and prediction of drug pharmacokinetics. J Control Release 352:961\u0026ndash;969\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRu J, Li P, Wang J et al (2014) TCMSP: a database of systems pharmacology for drug discovery from herbal medicines. J Cheminform 6:13\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang JF, Zhou H, Han LY et al (2005) TCM-ID: traditional Chinese medicine information database. Clin Pharmacol Ther 78(1):92\u0026ndash;93\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFang S, Dong L, Liu L et al (2021) HERB: a high-throughput experiment- and reference-guided database of traditional Chinese medicine. Nucleic Acids Res 49(D1):D1197\u0026ndash;D1206\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDaina A, Michielin O, Zoete V (2017) SwissADME:a free web tool to evaluate pharmacokinetics, drug-likeness, and medicinal chemistry friendliness of small molecules. Sci Rep 7:42717\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJung M, Wells D, Rusch J et al (2019) Unified single-cell analysis of testis gene regulation and pathology in five mouse strains. eLife. ;8\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBindea G, Mlecnik B, Tosolini M et al (2013) Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity 39(4):782\u0026ndash;795\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang C, Li C, Choi PY et al (2011) A novel method for molecular dynamics simulation in the isothermal\u0026ndash;isobaric ensemble. Mol Phys 109:191\u0026ndash;202\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHou T, Wang J, Li Y et al (2011) Assessing the performance of the MM/PBSA and MM/GBSA methods. 1. The accuracy of binding free energy calculations based on molecular dynamics simulations. J Chem Inf Model 51(1):69\u0026ndash;82\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRajapakse S, de Silva NL, Weeratunga P et al (2017) Prophylactic and therapeutic interventions for bleeding in dengue: a systematic review. Trans R Soc Trop Med Hyg 111:433\u0026ndash;439\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSuputtamongkol Y, Avirutnan P, Mairiang D et al (2021) Ivermectin accelerates circulating nonstructural protein 1 (NS1) clearance in adult dengue patients: A combined phase 2/3 randomized double-blinded placebo-controlled trial. Clin Infect Dis 72(10):e586\u0026ndash;e593\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilder-Smith A, Hombach J, Ferguson N et al (2019) Deliberations of the Strategic Advisory Group of Experts on Immunization on the use of CYD-TDV dengue vaccine. Lancet Infect Dis 19(1):e31\u0026ndash;e38\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKawai T, Akira S (2008) Toll-like receptor and RIG-I-like receptor signaling. Ann N Y Acad Sci 1143:1\u0026ndash;20\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKawai T, Akira S (2017) Signaling to NF-kappaB by Toll-like receptors. Trends Mol Med 13(11):460\u0026ndash;469\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChazal M, Beauclair G, Gracias S et al (2018) RIG-I recognizes the 5' region of Dengue and Zika virus genomes. Cell Rep 24(2):320\u0026ndash;328\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSolstad A, Hogaboam O, Forero A et al (2022) RIG-I-like receptor regulation of immune cell function and therapeutic implications. J Immunol 209(5):845\u0026ndash;854\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJusof FF, Lim CK, Aziz FN et al (2022) The cytokines CXCL10 and CCL2 and the kynurenine metabolite anthranilic acid accurately predict patients at risk of developing dengue with warning signs. J Infect Dis 226(11):1964\u0026ndash;1973\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJhan MK, HuangFu WC, Chen YF et al (2018) Anti-TNF-α restricts dengue virus-induced neuropathy. J Leukoc Biol 104(5):961\u0026ndash;968\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Zhang Q, Gui L et al (2018) Let-7e inhibits TNF-α expression by targeting the methyl transferase EZH2 in DENV2-infected THP-1 cells. J Cell Physiol 233(11):8605\u0026ndash;8616\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Boer ECW, van Gils JM, van Gils MJ (2020) Ephrin-Eph signaling usage by a variety of viruses. Pharmacol Res 159:105038\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLupberger J, Zeisel MB, Xiao F et al (2011) EGFR and EphA2 are host factors for hepatitis C virus entry and possible targets for antiviral therapy. Nat Med 17(5):589\u0026ndash;595\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"functional-and-integrative-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"fige","sideBox":"Learn more about [Functional \u0026 Integrative Genomics](http://link.springer.com/journal/10142)","snPcode":"10142","submissionUrl":"https://submission.nature.com/new-submission/10142/3","title":"Functional \u0026 Integrative Genomics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Dengue fever, Qingwen Baidu Decoction, Network pharmacology, Bioinformatic analysis, Machine learning","lastPublishedDoi":"10.21203/rs.3.rs-8520853/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8520853/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDengue fever, a mosquito-borne disease caused by the dengue virus (DENV), poses a serious global health challenge with no FDA-approved drug available. Qingwen Baidu Decoction (QBD) is a traditional Chinese medicine prescription used in China to treat critical dengue fever, though its mechanism remains unclear. This study aimed to elucidate the therapeutic mechanisms of QBD against dengue fever through a combination of network pharmacology, machine learning, and molecular docking. Potential targets of QBD were predicted using network pharmacology, and core genes were identified from DENV-upregulated genes via machine learning. Functional enrichment analyses, including GO and KEGG, were conducted to explore related biological processes and pathways. Single-cell RNA sequencing and immune infiltration analyses were performed to identify key cell subtypes and immune correlations. Molecular docking and dynamics simulations were used to evaluate interactions between QBD components and target proteins. Results indicated that QBD interferes with DENV entry by targeting EPHB2 and inhibits viral replication by binding structural protein E and C and nonstructural protein NS5. Additionally, QBD alleviates inflammatory responses by suppressing CXCL10 and EZH2, and modulates host immune responses during DENV infection.\u003c/p\u003e","manuscriptTitle":"Unveiling the molecular mechanism of Qing Wen Bai Du decoction against dengue fever: An integrated study of bioinformatic analysis, machine learning and network pharmacology","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-16 08:22:08","doi":"10.21203/rs.3.rs-8520853/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-18T09:10:19+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-18T03:21:31+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-17T12:46:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"306874378211346514776572006652979043213","date":"2026-03-17T12:40:31+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-17T02:52:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"173780234104144086698993784337388391137","date":"2026-03-17T02:27:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"113594606283166963108864781741515640812","date":"2026-03-16T07:50:43+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"28625242907170448776010913368445526033","date":"2026-03-13T10:36:26+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"229188881891648447959182679718732857097","date":"2026-03-11T13:58:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"228818119500867856977930248864275122747","date":"2026-03-11T13:47:40+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"286378076488331801677533470167618776268","date":"2026-03-11T13:47:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-11T13:42:24+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-09T12:22:13+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-09T12:21:50+00:00","index":"","fulltext":""},{"type":"submitted","content":"Functional \u0026 Integrative Genomics","date":"2026-01-05T11:19:20+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"functional-and-integrative-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"fige","sideBox":"Learn more about [Functional \u0026 Integrative Genomics](http://link.springer.com/journal/10142)","snPcode":"10142","submissionUrl":"https://submission.nature.com/new-submission/10142/3","title":"Functional \u0026 Integrative Genomics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"a312b22b-73e0-45dd-ad8b-e29c0bef73f5","owner":[],"postedDate":"March 16th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-10T00:53:32+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-16 08:22:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8520853","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8520853","identity":"rs-8520853","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}