Integrative multi-omics analysis identifies RPL5 and IL1R1 as potential diagnostic and therapeutic targets of sepsis

Integrative multi-omics analysis identifies RPL5 and IL1R1 as potential diagnostic and therapeutic targets of sepsis | 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 Integrative multi-omics analysis identifies RPL5 and IL1R1 as potential diagnostic and therapeutic targets of sepsis Mohan Li This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9624193/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Sepsis is a life-threatening condition characterized by dysregulated immune responses to infection. While advances in critical care have improved, sepsis-related mortality remains high, underscoring the urgent need for further research into therapeutic targets and potential candidate agents. Methods: We performed an integrative multi-omics analysis combining five bulk transcriptomic datasets and single-cell RNA sequencing data. Differential expression analysis and WGCNA were used to identify disease-associated modules. Machine learning algorithms, including LASSO, SVM-RFE, and random forest, were applied to screen key hub genes. Functional enrichment, immune infiltration analysis, and in silico gene perturbation were conducted to explore biological roles of these hub genes. An AI-driven drug prediction framework together with molecular docking were used to identify potential therapeutic compounds. Results: We identified 1,845 differentially expressed genes and sepsis-associated modules, among which the magenta and brown modules found by WGCNA analysis showed the strongest correlation with sepsis. Integrative analysis identified RPL5 and IL1R1 as key hub genes, with RPL5 demonstrating the highest diagnostic performance (AUC = 0.859). Functional analyses revealed that RPL5 is associated with ribosomal pathways and immune regulation, whereas IL1R1 is linked to inflammatory signaling. Single-cell analysis showed that RPL5 is broadly expressed across multiple immune cell types. In silico knockout indicated that RPL5 regulates immune-related pathways, including IL-17 signaling. Drug prediction identified BRD-K39768328 and BRD-K06569345 as candidate compounds with favorable binding affinities to both RPL5 and IL1R1. Conclusions: This study identified RPL5 and IL1R1 as potential diagnostic and therapeutic targets and proposed BRD-K39768328 and BRD-K06569345 as candidate compounds for sepsis, offering novel insights into sepsis pathogenesis and potential therapeutic interventions. Critical Care & Emergency Medicine Sepsis RPL5 IL1R1 Multi-omics integration Drug prediction Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Highlights Integrative multi-omics analysis identifies RPL5 and IL1R1 as key regulators in sepsis Single-cell and in silico knockout analyses reveal functional mechanisms AI-driven drug screening identifies potential therapeutic compounds for sepsis Introduction Sepsis is a life-threatening syndrome characterized by a dysregulated host response to infection [1]. It can progress to septic shock, the most severe form, which is associated with profound hypotension and multiple organ failure, substantially increasing the risk of mortality. Collectively, sepsis and septic shock remain leading causes of death worldwide. Although global sepsis-related mortality declined by 52.8% between 1990 and 2017, significant regional disparities persist, with the highest burden observed in low-income countries and areas [2]. Despite advances in supportive care, effective targeted therapies for sepsis remain limited, largely due to an incomplete understanding of its underlying molecular mechanisms [3]. Consistent with the definition of sepsis, it is considered a classical immune-driven disease, in which excessive inflammatory activation and subsequent immune suppression coexist and contribute to disease progression [4-6]. Among these, cytokine-mediated signaling pathways, such as interleukin-1 (IL-1) signaling, play central roles in initiating and amplifying inflammatory responses [7, 8]. Ribosome biogenesis is the process that generates ribosomes and plays an essential role in cell proliferation, differentiation, apoptosis, development, and transformation[9]. Under conditions of infection and inflammation, ribosomal function may be disrupted, leading to altered protein synthesis and impaired cellular function [10] . Notably, recent studies have identified ribosome-related pathways as significantly enriched in sepsis, with key genes associated with disease prognosis [11]. These findings suggest that dysregulation of translational control and ribosomal function may play a critical role in modulating immune responses during sepsis. With the rapid development of high-throughput sequencing technologies, integrative analysis of bulk RNA sequencing and single-cell RNA sequencing (scRNA-seq) has enabled a more comprehensive understanding of disease-associated molecular networks and cellular heterogeneity. In parallel, machine learning approaches have improved the identification of robust biomarkers from complex datasets. However, integrative multi-omics studies combined with computational modeling to identify targetable genes and therapeutic candidates in sepsis remain scarce. In this study, we performed an integrative multi-omics analysis combining bulk transcriptomics, single-cell RNA sequencing, and machine learning approaches to systematically identify key regulators in sepsis. Moreover, AI-driven drug screening and molecular docking analyses were used to identify potential therapeutic compounds targeting these hub genes. These findings provide new insights into the molecular mechanisms of sepsis and suggest novel directions for therapeutic intervention. Materials and methods Source of data Five sepsis-related whole-blood microarray datasets (GSE13015, GSE137342, GSE236713, GSE54514, and GSE69063) along with clinical information were obtained from the Gene Expression Omnibus (GEO) database using the R package ‘ GEOquery’ [12]. These datasets were subsequently integrated, and batch effects were corrected by using the ‘ sva ’ package. Data normalization and standardization were performed using the ‘ limma ’ package afterwards [13]. Identification of DEGs and WGCNA analysis Differential expression analysis was performed by utilizing the ‘ limma ’ R package [13]. Differentially expressed genes (DEGs) were then identified with thresholds of |log 2 fold change (FC)| > 1 and p < 0.05, and visualized using volcano plots and heatmaps generated by applying the ‘ ggplot2 ’ and ‘ ComplexHeatmap ’ packages [14]. Weighted gene co-expression network analysis (WGCNA) was conducted using the ‘ WGCNA ’ package [15] to further explore gene co-expression patterns. Genes were clustered into modules using average linkage hierarchical clustering and modules with high similarity were merged based on module eigengene correlation. Module–trait relationships were then assessed to find out modules which are significantly associated with sepsis. Hub genes were identified by intersecting DEGs with genes from key co-expression modules, and the results were visualized using Venn diagrams. Functional enrichment analyses, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, were performed using the ‘ clusterProfiler ’ package [16] with a false discovery rate (FDR) < 0.05. Machine learning algorithms and diagnostic model construction Multiple machine learning algorithms were applied to identify hub genes, including least absolute shrinkage and selection operator (LASSO), support vector machine-recursive feature elimination (SVM-RFE), and random forest (RF). LASSO logistic regression was used to select variables by shrinking regression coefficients with an L1 penalty, thereby reducing less informative features to zero. SVM-RFE was employed to iteratively eliminate features and identify the most relevant genes. RF analysis was performed to evaluate the importance of each variable based on decision tree models. Hub genes were determined by integrating the results of these three algorithms and visualized by Venn diagram. The expression patterns and diagnostic performance of hub genes were validated across different datasets. Receiver operating characteristic (ROC) curves and nomograms were generated using the ‘ pROC ’, ‘ rms ’, and ‘ rmda ’ R packages to evaluate diagnostic efficacy of the hub genes found based on machine learning. To further explore the functional roles of hub genes, single-gene gene set enrichment analysis (GSEA) was performed using the ‘ clusterProfiler ’ package [16] based on the hallmark gene sets obtained from the MSigDB database[17]. Then, Immune cell infiltration was estimated using the ‘CIBERSORT ‘package[18] of R. Single-cell transcriptomic analysis The processed single-cell RNA sequencing (scRNA-seq) dataset (GSE216009) was also obtained from the Gene Expression Omnibus (GEO) database. Given that the data were provided as a preprocessed ‘Seurat’ object, no additional raw data preprocessing or quality control was performed. Downstream analyses were conducted using the ‘ Seurat ’ R package. The expression patterns of hub genes were evaluated across different cell populations. In addition, gene perturbation analysis was performed using ‘ scTenifoldKnk ’ R package [19] to assess the impact of hub gene knockout. AI-driven drug prediction and molecular docking Integrated bulk RNA-seq datasets were used as input to screen for potential drugs capable of modulating disease-associated gene expression patterns. The DrugRefLector [20] framework was applied to identify candidate therapeutic compounds based on transcriptomic profiles. Molecular docking analysis was performed between candidate compounds and hub gene-encoded protein to further evaluate drug–target interactions. Protein structures were obtained from the RCSB Protein Data Bank (PDB) [21], and ligand structures were retrieved from the PubChem database [22]. Protein structures were preprocessed using PyMOL to remove water molecules and bound ligands. Molecular docking was conducted using AutoDock Vina to predict binding affinities and identify potential binding sites. Docking scores (kcal/mol) were calculated for each protein–ligand interaction, and the conformation with the lowest binding energy was selected as the optimal binding mode. The docking results were visualized using PyMOL to illustrate binding interactions, including hydrogen bond formation between ligands and target proteins. Statistical analysis All statistical analyses were performed in R software (Version 4.5.3). Differences between two groups were assessed using Student’s t-test. A two-tailed p-value < 0.05 was considered statistically significant. Results The study design is presented in Figure 1A. First of all, we integrated five bulk RNA-seq datasets from independent studies to construct a unified transcriptomic cohort for following analyses. Prior to batch correction, principal component analysis (PCA) revealed strong dataset-driven clustering, with samples segregating primarily according to their study of origin rather than biological differences between sepsis and healthy individuals (Figure 1B). This indicates the presence of substantial batch effects across datasets. Thereby, batch correction was performed to mitigate these confounding effects. Following correction, PCA demonstrated a marked reduction in dataset-specific clustering, with samples from different cohorts becoming more intermixed (Figure 1C). Consistently, global gene expression distributions across samples became more comparable after batch correction, as shown by the aligned expression profiles across datasets (Figure 1D), enabling reliable downstream integrative analyses. PCA (Figure 2A) revealed a significant separation between sepsis patients and healthy individuals along PC1 (p < 0.001), despite partial overlap between groups. A total of 1,845 DEGs were identified, including 599 upregulated and 1246 downregulated genes (Figure 2B). Distinct expression profiles between patients with sepsis and healthy individuals was further illustrated by heatmap (Figure 2C). Multiple gene co-expression modules across the dataset were found by WGCNA (Figure 2D). Module–trait relationship analysis revealed several modules significantly associated with disease status (Figure 2E). In particular, the brown module showed strong positive correlation with disease, while the magenta module exhibited a negative association (Figure 2E). Consistent with the module–trait correlation analysis (Figure 2E), the brown and magenta modules also exhibited the highest gene significance values (Figure 2F), further confirming their strong association with disease status. A total of 4,178 genes were identified within the magenta module, of which 1,216 genes overlapped with differentially expressed genes (DEGs), as shown in the Venn diagram (Figure 3A). KEGG pathway enrichment analysis revealed that these overlapping genes were predominantly enriched in pathways associated with the ribosome, human T-cell leukemia virus 1 infection, coronavirus disease (COVID-19), and cytokine–cytokine receptor interaction (Figure 3B). GO enrichment analysis further demonstrated that these genes were involved in immune-related biological processes (BP), including lymphocyte differentiation, leukocyte cell–cell adhesion, and regulation of T cell activation. In terms of cellular components (CC), these genes were mainly enriched in ribosome-associated structures. Moreover, molecular function (MF) analysis indicated enrichment in catalytic activity, structural constituent of ribosome, and immune receptor activity (Figure 3C). Similarly, 1,268 genes were identified in the brown module, among which 440 genes overlapped with DEGs (Figure S1A). KEGG pathway analysis indicated that these genes were primarily enriched in cytokine–cytokine receptor interaction and the MAPK signaling pathway (Figure S1B). GO enrichment analysis further revealed their significant enrichment in immune-related biological processes, such as regulation of immune effector processes, inflammatory response, and leukocyte-mediated immunity. In addition, molecular function analysis highlighted enrichment in immune receptor activity (Figure S1C). Taken together, these findings suggest that the magenta module integrates both immune and translational processes, whereas the brown module predominantly represents canonical inflammatory signaling, highlighting the critical roles of translational dysregulation and inflammation in sepsis. The 1,216 overlapping genes (Figure 3A) were further analyzed using a protein–protein interaction (PPI) network, revealing a densely interconnected network architecture (Figure 4A). Based on node connectivity, the top 10 hub genes were identified according to their degree centrality within the network (Figure 4B). To further assess candidate biomarkers, these hub genes were subjected to multiple machine learning approaches, including LASSO regression, SVM-RFE, and random forest analysis (Figure 4C–H). Integrative analysis of these three algorithms identified five overlapping key genes—RPL5, RPS23, RPS27A, RPS19, and RPS3A (Figure 4I). Notably, all identified genes are ribosome-associated proteins, further supporting the involvement of translational regulation in sepsis pathogenesis. A diagnostic nomogram based on these five genes was subsequently constructed to distinguish sepsis patients from healthy controls (Figure 4J). ROC curves demonstrated robust diagnostic performance of these 5 genes, among which RPL5 exhibited the highest predictive accuracy, with an area under the curve (AUC) of 0.859 (Figure 4K). Similarly, the 440 overlapping genes derived from the brown module (Figure S1A) were subjected to PPI network analysis, revealing a complex interaction network (Figure S2A). The top 10 hub genes were identified based on node connectivity (Figure S2B) and subsequently analyzed using LASSO regression, SVM-RFE, and random forest approaches (Figure S2C–H). Integrative analysis identified seven overlapping key genes—JAK2, FCER1G, IL1R1, JAK3, MAPK14, VSIG4, and CSF3R (Figure S2I), which are predominantly involved in classical inflammatory signaling pathways. A diagnostic nomogram based on these genes was constructed (Figure S2J), and ROC analysis demonstrated strong predictive performance of these key genes, with IL1R1 showing the highest diagnostic accuracy (AUC = 0.855) (Figure S2K). Collectively, these findings further support a dual regulatory framework in sepsis, in which ribosome-associated genes (e.g., RPL5) represent the translational arm, whereas immune signaling genes (e.g., IL1R1) reflect the inflammatory arm, together contributing to the pathogenesis of sepsis. GSEA was applied to further elucidate the molecular role of RPL5. Samples with high RPL5 expression were significantly enriched in multiple immune-related pathways, including cytokine–cytokine receptor interaction and T cell–associated signaling pathways (Figure 5A). Samples with low RPL5 expression were enriched in complement and coagulation cascades, as well as other immune-related pathways (Figure 5B). CIBERSORT analysis was conducted to further characterize the immune microenvironment associated with RPL5, and the results showed that RPL5 expression was significantly correlated with the proportions of several immune cell populations, including CD4 memory T cells (resting and activated), regulatory T cells (Tregs), and monocytes. Notably, RPL5 expression was negatively correlated with macrophage subsets (M0, M1, and M2) (Figure 5C), indicating that reduced RPL5 expression may be associated with enhanced macrophage-driven inflammatory responses and impaired adaptive immune regulation. Similarly, GSEA was performed to investigate the molecular role of IL1R1. It was observed that samples with high IL1R1 expression were significantly enriched in classical inflammatory signaling pathways, including MAPK signaling and Toll-like receptor signaling pathways (Figure S3A). In contrast, samples with low IL1R1 expression showed enrichment in pathways such as allograft rejection and asthma (Figure S3B). Consistent with these findings, CIBERSORT analysis revealed that IL1R1 expression was significantly associated with multiple immune cell populations, including memory B cells, CD8 T cells, monocytes, macrophages (M0 and M2), and neutrophils (Figure S3C). Single-cell RNA sequencing analysis was performed using the GSE216009 dataset to further validate the role of RPL5 at single-cell resolution, which included 12 healthy individuals and 52 sepsis patients. UMAP visualization identified 24 distinct cell types across all samples (Figure 6A), with neutrophils and T cells representing the most abundant populations. The expression pattern of RPL5 was subsequently examined, revealing that it is broadly expressed across multiple immune cell types, particularly in T cells, monocytes, plasmacytoid dendritic cells (pDCs), and hematopoietic stem and progenitor cells (HSPCs) (Figure 6B–C), suggesting a potential role in coordinating immune cell function. To further understand the functional impact of RPL5, in silico knockout analysis was performed. Approximately 0.5% of genes were significantly differentially expressed following RPL5 knockout (Figure 6D). The top 20 DEGs included ANXA3, CEACAM8, STOM, CD24, and UGCG (Figure 6E). Functional enrichment analysis demonstrated that these DEGs were significantly involved in immune-related biological processes, such as defense response to bacterium and cell killing (Figure 6F). Consistently, KEGG pathway analysis revealed enrichment in immune-associated pathways, particularly the IL-17 signaling pathway (Figure 6G), highlighting a potential role for RPL5 in modulating host immune responses during sepsis. Likewise, to assess the functional significance of IL1R1, in silico knockout analysis was conducted. Approximately 0.5% of genes were significantly differentially expressed following IL1R1 knockout (Figure S4A). The top 20 DEGs overlapped substantially with those observed in RPL5 knockout, including ANXA3, CEACAM8, CD24, STOM, and UGCG (Figure S4B). Functional enrichment analysis showed that these DEGs were significantly involved in immune-related biological processes, including defense response to bacterium, defense response to Gram-negative bacterium, and antimicrobial humoral response (Figure S4C). KEGG pathway analysis found their enrichment in cobalamin transport and metabolism (Figure S4D), suggesting additional metabolic regulation associated with IL1R1 signaling. We applied an AI-driven drug screening framework (DrugRefLector) using an integrated bulk transcriptomic profile derived from five microarray datasets. This analysis identified ten candidate compounds capable of reversing sepsis-associated transcriptional signatures (Figure 7A), among which BRD-K39768328 (Figure 7B) and BRD-K06569345 (Figure 7C) emerged as the top-ranked candidates. To assess whether RPL5 could serve as a direct target of these compounds, molecular docking analysis was performed. The results suggested that BRD-K39768328 binds to the C1 cavity pocket of RPL5 with a binding affinity of −7.3 kcal/mol (Figure 7D), while BRD-K06569345 binds to the C4 cavity pocket with a binding affinity of −7.5 kcal/mol (Figure 7E), indicating stable and favorable interactions. Molecular docking analysis also provided the evidence that BRD-K39768328 binds to the C1 cavity pocket of IL1R1 with a binding affinity of −7.2 kcal/mol (Figure S5A), whereas BRD-K06569345 binds to the C2 cavity pocket with an even stronger binding affinity of −8.3 kcal/mol (Figure S5B), suggesting high binding stability. Notably, both compounds exhibited binding affinity toward RPL5 and IL1R1, indicating a potential dual-target therapeutic mechanism. Discussion In this study, we performed an integrative multi-omics analysis combining bulk transcriptomics, single-cell RNA sequencing, and machine learning approaches to uncover key regulatory mechanisms in sepsis. Our results identified RPL5 and IL1R1 as two central hub genes, representing the translational regulation and inflammatory signaling axis, respectively. These findings suggest that simultaneous targeting of these two pathways may represent a promising therapeutic strategy for sepsis. Therefore, we applied an AI-driven drug screening framework and identified two candidate compounds, BRD-K39768328 and BRD-K06569345. Molecular docking analysis demonstrated that both compounds exhibit stable binding affinities toward RPL5 and IL1R1, supporting their potential as dual-target therapeutic agents. Accumulating evidence suggests that ribosomal proteins are not only structural components of the translational machinery but also play critical roles in immune regulation. For instance, ribosomal protein S3 (RPS3) serves as a critical regulator in the NF-κB signaling pathway, playing an essential role in determining promoter selectivity and transcriptional specificity of NF-κB [ 23 ]. Another study has showed that silencing RPS3 can mitigate cigarette smoke-induced acute lung injury by inhibiting NF-κB activity [ 24 ]. Importantly, RPS3 has been reported to participate in the inflammatory response in sepsis-associated acute kidney injury (S-AKI) through activation of the NF-κB signaling pathway [ 25 ]. In aligning with these findings, we also observed that ribosomal protein S3A (RPS3A) was a key gene within the magenta module and, together with other ribosome-associated genes, exhibited strong diagnostic value in distinguishing patients with sepsis from healthy individuals. Beyond RPS3, growing evidence indicates that other ribosome-associated genes are frequently identified as key regulators in sepsis-related conditions. Previous studies have demonstrated that several ribosomal genes serve as central hub genes in sepsis-related myopathy, suggesting that ribosome-associated processes may be broadly involved in immune dysregulation during sepsis [ 26 ]. Building upon these evidence, our study extends the role of ribosomal proteins to ribosomal protein L5 (RPL5), a ribosomal protein involved in ribosome biogenesis and translational control, which has been largely unexplored in the context of sepsis. In this study, RPL5 was found to be significantly downregulated in sepsis and was associated with ribosome-related pathways and immune processes. Its downregulation may impair ribosome function and disrupt translational capacity. Notably, we found that RPL5 was broadly expressed across multiple immune cell types, including T cells and monocytes, implying its potential role in maintaining immune cell function. Single-cell transcriptomic analysis further supported these findings by revealing cell-type-specific expression patterns of RPL5 and its involvement in immune cell populations. In silico gene perturbation analysis demonstrated that RPL5 knockout affected a functionally relevant subset of genes, particularly those involved in immune-related pathways such as IL-17 signaling, indicating a targeted regulatory role rather than a global transcriptional effect. Unlike RPS3, which has been shown to actively promote NF-κB signaling, our results indicate that RPL5 is significantly downregulated in sepsis, implying a distinct mechanism. We hypothesize that reduced RPL5 expression may impair ribosomal integrity and translational capacity, thereby contributing to immune dysregulation. Collectively, these findings highlight a broader role of ribosomal proteins in immune regulation and support the notion that dysregulation of ribosome-associated processes may represent an important mechanism underlying sepsis pathogenesis which warrant further investigation. Interleukin-1 receptor type 1 (IL1R1) is a critical cytokine receptor belonging to the IL-1 receptor family, primarily acting as a high-affinity receptor for proinflammatory cytokines IL-1α and IL-1β. It mediates immune and inflammatory responses, including neutrophil recruitment and NF-𝜅B activation [ 27 – 29 ]. Consistent with previous studies, our results demonstrated that IL1R1 is significantly upregulated in sepsis and is predominantly expressed in inflammatory immune cell populations, including monocytes, macrophages, and neutrophils. These findings support the role of IL1R1 as a key driver of immune dysregulation and cytokine storm in sepsis. We also found that IL1R1 exhibited strong diagnostic performance in distinguishing patients with sepsis from healthy individuals, indicating that it may serve not only as a potential therapeutic target but also as a clinically relevant diagnostic biomarker. From a translational perspective, our work is the first study that identified plausible therapeutic compounds using an AI-driven drug prediction framework in the context of sepsis. Notably, BRD-K39768328 and BRD-K06569345 showed favorable binding affinities with both RPL5 and IL1R1, highlighting their potential to simultaneously modulate ribosome-associated and inflammatory pathways. This dual-target strategy may represent a promising direction for future therapeutic development in sepsis. However, it must be noted that several limitations should be acknowledged. First, our study is primarily based on publicly available datasets and computational analyses, lacking experimental validation. Second, the mechanistic interactions between IL1R1 and RPL5 require further investigation in vitro and in vivo. Future studies are needed to validate these findings and explore their clinical applicability. In conclusion, our study identifies RPL5 and IL1R1 as reliable diagnostic biomarkers and therapeutic targets in sepsis. In addition, BRD-K39768328 and BRD-K06569345 were identified as promising candidate compounds with the potential to reverse sepsis-associated gene expression patterns, warranting further validation in preclinical models. These findings provide new insights into the pathogenesis of sepsis and offer potential avenues for the development of targeted therapeutic strategies. Declarations Declaration of competing interests: The authors state that they have no conflicts of interest, as per the journal's definition. Ethical Approval and Consent to participate: This study was conducted based entirely on publicly available datasets retrieved from the Gene Expression Omnibus (GEO) database. All data used in this study were previously published and approved by the respective institutional review boards. Therefore, no additional ethical approval or informed consent was required for this study. Consent for publication: Not applicable. Availability of supporting data: The datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/). All code used in this study is available upon request. Competing interests: The authors declare that they have no competing interests. Funding: None. Authors' contributions: M.L. conceived and designed the study. M.L. conducted statistical analyses and data visualization. M.L. wrote the original draft of the manuscript. Acknowledgements: The authors would like to thank the contributors of the Gene Expression Omnibus (GEO) database for providing publicly available datasets, which made this study possible. References Singer, M., et al., The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Jama, 2016. 315 (8): p. 801–10. Rudd, K.E., et al., Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet, 2020. 395 (10219): p. 200–211. Li, M., et al., High-density lipoprotein: a biomarker and therapeutic target in sepsis. Crit Care, 2025. 29 (1): p. 453. Wiersinga, W.J., et al., Host innate immune responses to sepsis. Virulence, 2014. 5 (1): p. 36–44. Willmann, K. and L.F. Moita, Physiologic disruption and metabolic reprogramming in infection and sepsis. Cell Metab, 2024. 36 (5): p. 927–946. Cao, M., G. Wang, and J. Xie, Immune dysregulation in sepsis: experiences, lessons and perspectives. Cell Death Discov, 2023. 9 (1): p. 465. Zhang, Y.Y. and B.T. Ning, Signaling pathways and intervention therapies in sepsis. Signal Transduct Target Ther, 2021. 6 (1): p. 407. Weber, A., P. Wasiliew, and M. Kracht, Interleukin-1 (IL-1) pathway. Sci Signal, 2010. 3 (105): p. cm1. Jiao, L., et al., Ribosome biogenesis in disease: new players and therapeutic targets. Signal Transduct Target Ther, 2023. 8 (1): p. 15. Sundaramoorthy, E., et al., Ribosome quality control activity potentiates vaccinia virus protein synthesis during infection. J Cell Sci, 2021. 134 (8). Wang, H., et al., Preliminary screening of new biomarkers for sepsis using bioinformatics and experimental validation. PLoS One, 2025. 20 (1): p. e0317608. Davis, S. and P.S. Meltzer, GEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor. Bioinformatics, 2007. 23 (14): p. 1846–7. Ritchie, M.E., et al., limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res, 2015. 43 (7): p. e47. Gu, Z., Complex heatmap visualization. Imeta, 2022. 1 (3): p. e43. Langfelder, P. and S. Horvath, WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics, 2008. 9 : p. 559. Yu, G., et al., clusterProfiler: an R package for comparing biological themes among gene clusters. Omics, 2012. 16 (5): p. 284–7. Liberzon, A., et al., The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst, 2015. 1 (6): p. 417–425. Chen, B., et al., Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol, 2018. 1711 : p. 243–259. Osorio, D., et al., scTenifoldKnk: An efficient virtual knockout tool for gene function predictions via single-cell gene regulatory network perturbation. Patterns (N Y), 2022. 3 (3): p. 100434. DeMeo, B., et al., Active learning framework leveraging transcriptomics identifies modulators of disease phenotypes. Science, 2025. 390 (6776): p. eadi8577. Berman, H.M., et al., The Protein Data Bank. Nucleic Acids Res, 2000. 28 (1): p. 235–42. Kim, S., Getting the most out of PubChem for virtual screening. Expert Opin Drug Discov, 2016. 11 (9): p. 843–55. Wan, F., et al., IKKβ phosphorylation regulates RPS3 nuclear translocation and NF-κB function during infection with Escherichia coli strain O157:H7. Nat Immunol, 2011. 12 (4): p. 335–43. Dong, J., et al., Ribosomal Protein S3 Gene Silencing Protects Against Cigarette Smoke-Induced Acute Lung Injury. Mol Ther Nucleic Acids, 2018. 12 : p. 370–380. Zhang, X., et al., RPS3 Aggravates Sepsis-Induced Acute Kidney Injury Through Activating NF-kappaB Mediated Renal Inflammatory Responses. Physiol Res, 2026. 75 (1): p. 63–77. Wang, J., K. Han, and J. Lu, Screening of hub genes for sepsis-induced myopathy by weighted gene co-expression network analysis and protein-protein interaction network construction. BMC Musculoskelet Disord, 2024. 25 (1): p. 834. Ralph, B.A., et al., The IL-1 Receptor Is Required to Maintain Neutrophil Viability and Function During Aspergillus fumigatus Airway Infection. Front Immunol, 2021. 12 : p. 675294. Guo, Q., et al., NF-κB in biology and targeted therapy: new insights and translational implications. Signal Transduct Target Ther, 2024. 9 (1): p. 53. Gehrke, N., et al., Hepatocyte-specific deletion of IL1-RI attenuates liver injury by blocking IL-1 driven autoinflammation. J Hepatol, 2018. 68 (5): p. 986–995. Additional Declarations The authors declare no competing interests. Supplementary Files FigureS1toS5.docx Graphicalabstract.docx Cite Share Download PDF Status: Posted Version 1 posted 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-9624193","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":635146606,"identity":"e0f13c20-eb5f-4e06-879c-7a1b3f2210b1","order_by":0,"name":"Mohan Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzElEQVRIiWNgGAWjYHACNhAhw8beAKQKGBgkiNXCw8ZzgIHhgAEpWhgkEojUotvewPbg5w4GHj7Jx8ekPxjYyEk2MD98dAOPFrMzB9gNe88AHSadliZxwCDNWJqBzdg4B5+WGwlsErxtIC05ZkAthxPngdkEtEj+BWmRPEOCFmmwLRI8EC2zCWo5c7BNWrZNAhjIackWZ4B+kWwm5Jfjzcck37bZyMm3Hz54o6LCRk7iePPDx/i0MDAwNjCgRgYzXuWjYBSMglEwCogBAN0/PdVdbbilAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-5472-3699","institution":"Nanjing Drum Tower Hospital","correspondingAuthor":true,"prefix":"","firstName":"Mohan","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2026-05-06 03:19:51","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9624193/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9624193/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108677441,"identity":"c1cb8e27-db1e-4ce3-8b12-5aeb893cbbab","added_by":"auto","created_at":"2026-05-07 08:50:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":217315,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eData integration and batch effect correction across multiple transcriptomic cohorts.\u003c/strong\u003e(A) Schematic overview of the study workflow. (B) Principal component analysis (PCA) before batch correction. (C) PCA after batch correction. (D) Distribution of gene expression across samples after batch correction.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/10d3eecff359b91796c7ee6e.png"},{"id":108677442,"identity":"9ee7aaf5-79a4-4b25-a751-a0cf49240193","added_by":"auto","created_at":"2026-05-07 08:50:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":411952,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of disease-associated gene modules using differential expression and WGCNA. \u003c/strong\u003e(A) PCA of samples after batch correction colored by disease status (Control vs Sepsis). (B) Volcano plot showing differentially expressed genes (DEGs) between sepsis patients and healthy controls. Upregulated and downregulated genes are highlighted based on log2 fold change and statistical significance. (C) Heatmap of differentially expressed genes across samples. (D) Gene dendrogram and module assignment identified by WGCNA. (E) Correlations between module eigengenes and clinical traits are shown.(F) Gene significance across modules.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/14955afa46220cfafbd27e3d.png"},{"id":108806137,"identity":"e3058c45-6340-49d8-a859-2c07120a8ac0","added_by":"auto","created_at":"2026-05-08 15:27:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":210924,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunctional characterization of intersecting genes between the magenta module and differentially expressed genes.\u003c/strong\u003e (A) Venn diagram showing the overlap between genes in the magenta module identified by WGCNA and differentially expressed genes (DEGs). (B) KEGG pathway enrichment analysis of the overlapping genes. (C) Gene Ontology (GO) enrichment analysis of the overlapping genes. Biological process (BP), cellular component (CC), and molecular function (MF) categories are shown.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/ccd1c48ddc5f77c3b26cecb7.png"},{"id":108677445,"identity":"3aaf6a4f-61b1-42a1-bc56-7b6ca5fc627c","added_by":"auto","created_at":"2026-05-07 08:50:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":274549,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of key hub genes and construction of a diagnostic model using integrative machine learning approaches based on magenta module. \u003c/strong\u003e(A) Protein–protein interaction (PPI) network of the 1,216 overlapping genes derived from the magenta module and DEGs. (B) Top 10 hub genes ranked by node degree (number of adjacent connections) in the PPI network. (C–D) Feature selection using LASSO regression. (C) LASSO coefficient profiles of candidate genes. (D) Cross-validation plot showing the optimal penalty parameter (λ). (E–H) Feature selection using random forest analysis. (E–F) Cross-validation curves for determining the optimal number of variables. (G) Error rate of the random forest model across different numbers of trees. (H) Variable importance ranking of candidate genes. (I) Venn diagram showing the overlap of selected genes from LASSO, random forest, and SVM-RFE analyses. (J) Nomogram constructed based on the five selected hub genes for predicting sepsis risk. (K) Receiver operating characteristic (ROC) curves of the five hub genes, showing their diagnostic performance.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/c2c359daa66a69469b4d1610.png"},{"id":108677446,"identity":"ae37f527-c0d8-4194-92a1-3c88e147d75b","added_by":"auto","created_at":"2026-05-07 08:50:15","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":186893,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunctional and immune landscape associated with RPL5 expression in sepsis.\u003c/strong\u003e (A) Gene set enrichment analysis (GSEA) of samples with high RPL5 expression. (B) GSEA of samples with low RPL5 expression. (C) Immune cell infiltration analysis using CIBERSORT. Comparison of immune cell fractions between high and low RPL5 expression groups.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/258d66998b83c6e7512f10cd.png"},{"id":108805403,"identity":"c5034f46-073c-4c8e-82f2-2de5ce88d73c","added_by":"auto","created_at":"2026-05-08 15:25:52","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":213132,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSingle-cell transcriptomic characterization and functional impact of RPL5 in sepsis.\u003c/strong\u003e (A) UMAP visualization of single-cell RNA-seq data from GSE216009, showing 24 annotated cell types across healthy controls and sepsis patients. (B) Dot plot showing the expression patterns of canonical marker genes across identified cell types. (C) UMAP feature plot displaying the expression distribution of RPL5 across different cell populations. (D) Proportion of differentially expressed genes following in silico RPL5 knockout. (E) Top 20 differentially expressed genes (DEGs) following RPL5 knockout. (F) Gene Ontology (GO) enrichment analysis of the top DEGs. (G) KEGG pathway enrichment analysis of the top DEGs.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/c93c7d5b5a82be4d37f7b84b.png"},{"id":108677448,"identity":"1929e560-8d15-4ec6-82a6-d4bf797269e6","added_by":"auto","created_at":"2026-05-07 08:50:15","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":306173,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of potential therapeutic agents for sepsis treatment targeting RPL5. \u003c/strong\u003e(A) Candidate drug screening results from DrugRefLector prediction.(B-C) 2D chemical structure depiction of (B) BRD-K39768328 and (C) BRD-K06569345. (D-E) Molecular docking between (D) BRD-K39768328 and (E) BRD-K06569345 and RPL5.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/2680a268b6b0802b783e3725.png"},{"id":108809612,"identity":"246f7283-25b0-494a-ba60-c71100f1985e","added_by":"auto","created_at":"2026-05-08 15:54:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2075523,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/281c42de-48d4-4356-8d89-4f1b520541e1.pdf"},{"id":108677443,"identity":"d21bf889-e62b-4c00-8afe-5a4eb1f7e731","added_by":"auto","created_at":"2026-05-07 08:50:15","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1089461,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS1toS5.docx","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/6ac77e477f71c44e48cdbc90.docx"},{"id":108677449,"identity":"5933fb6e-7ceb-4610-9c2a-c92d645e0dfc","added_by":"auto","created_at":"2026-05-07 08:50:16","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":413336,"visible":true,"origin":"","legend":"","description":"","filename":"Graphicalabstract.docx","url":"https://assets-eu.researchsquare.com/files/rs-9624193/v1/ce787f469033423b982ece12.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eIntegrative multi-omics analysis identifies RPL5 and IL1R1 as potential diagnostic and therapeutic targets of sepsis\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Highlights","content":"\u003cul\u003e\n \u003cli\u003eIntegrative multi-omics analysis identifies RPL5 and IL1R1 as key regulators in sepsis\u003c/li\u003e\n \u003cli\u003eSingle-cell and in silico knockout analyses reveal functional mechanisms\u003c/li\u003e\n \u003cli\u003eAI-driven drug screening identifies potential therapeutic compounds for sepsis\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"Introduction","content":"\u003cp\u003eSepsis is a life-threatening syndrome characterized by a dysregulated host response to infection [1]. It can progress to septic shock, the most severe form, which is associated with profound hypotension and multiple organ failure, substantially increasing the risk of mortality. Collectively, sepsis and septic shock remain leading causes of death worldwide. Although global sepsis-related mortality declined by 52.8% between 1990 and 2017, significant regional disparities persist, with the highest burden observed in low-income countries and areas [2].\u0026nbsp;Despite advances in supportive care, effective targeted therapies for sepsis remain limited, largely due to an incomplete understanding of its underlying molecular mechanisms [3].\u003c/p\u003e\n\u003cp\u003eConsistent with the definition of sepsis, it is considered a classical immune-driven disease, in which excessive inflammatory activation and subsequent immune suppression coexist and contribute to disease progression [4-6]. Among these, cytokine-mediated signaling pathways, such as interleukin-1 (IL-1) signaling, play central roles in initiating and amplifying inflammatory responses [7, 8]. Ribosome biogenesis is the process that generates ribosomes and plays an essential role in cell proliferation, differentiation, apoptosis, development, and transformation[9]. Under conditions of infection and inflammation, ribosomal function may be disrupted, leading to altered protein synthesis and impaired cellular function [10] . Notably, recent studies have identified ribosome-related pathways as significantly enriched in sepsis, with key genes associated with disease prognosis [11]. These findings suggest that dysregulation of translational control and ribosomal function may play a critical role in modulating immune responses during sepsis.\u003c/p\u003e\n\u003cp\u003eWith the rapid development of high-throughput sequencing technologies, integrative analysis of bulk RNA sequencing and single-cell RNA sequencing (scRNA-seq) has enabled a more comprehensive understanding of disease-associated molecular networks and cellular heterogeneity. In parallel, machine learning approaches have improved the identification of robust biomarkers from complex datasets. However, integrative multi-omics studies combined with computational modeling to identify targetable genes and therapeutic candidates in sepsis remain scarce.\u003c/p\u003e\n\u003cp\u003eIn this study, we performed an integrative multi-omics analysis combining bulk transcriptomics, single-cell RNA sequencing, and machine learning approaches to systematically identify key regulators in sepsis. Moreover, AI-driven drug screening and molecular docking analyses were used to identify potential therapeutic compounds targeting these hub genes. These findings provide new insights into the molecular mechanisms of sepsis and suggest novel directions for therapeutic intervention.\u003cstrong\u003e\u003c/strong\u003e\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003e\u003cstrong\u003eSource of data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFive sepsis-related whole-blood microarray datasets (GSE13015, GSE137342, GSE236713, GSE54514, and GSE69063) along with clinical information were obtained from the Gene Expression Omnibus (GEO) database using the R package \u0026lsquo;\u003cem\u003eGEOquery\u0026rsquo;\u0026nbsp;\u003c/em\u003e[12]. These datasets were subsequently integrated, and batch effects were corrected by using the \u0026lsquo;\u003cem\u003esva\u003c/em\u003e\u0026rsquo; package. Data normalization and standardization were performed using the \u0026lsquo;\u003cem\u003elimma\u003c/em\u003e\u0026rsquo; package afterwards [13].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIdentification of DEGs and WGCNA analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDifferential expression analysis was performed by utilizing the \u0026lsquo;\u003cem\u003elimma\u003c/em\u003e\u0026rsquo; R package [13]. Differentially expressed genes (DEGs) were then identified with thresholds of |log\u003csub\u003e2\u003c/sub\u003e fold change (FC)| \u0026gt; 1 and p \u0026lt; 0.05, and visualized using volcano plots and heatmaps generated by applying the \u0026lsquo;\u003cem\u003eggplot2\u003c/em\u003e\u0026rsquo; and \u0026lsquo;\u003cem\u003eComplexHeatmap\u003c/em\u003e\u0026rsquo; packages [14]. Weighted gene co-expression network analysis (WGCNA) was conducted using the \u0026lsquo;\u003cem\u003eWGCNA\u003c/em\u003e\u0026rsquo; package [15] to further explore gene co-expression patterns. Genes were clustered into modules using average linkage hierarchical clustering and modules with high similarity were merged based on module eigengene correlation. Module\u0026ndash;trait relationships were then assessed to find out modules which are significantly associated with sepsis. Hub genes were identified by intersecting DEGs with genes from key co-expression modules, and the results were visualized using Venn diagrams. Functional enrichment analyses, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, were performed using the \u0026lsquo;\u003cem\u003eclusterProfiler\u003c/em\u003e\u0026rsquo; package [16] with a false discovery rate (FDR) \u0026lt; 0.05.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMachine learning algorithms and diagnostic model construction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMultiple machine learning algorithms were applied to identify hub genes, including least absolute shrinkage and selection operator (LASSO), support vector machine-recursive feature elimination (SVM-RFE), and random forest (RF).\u003c/p\u003e\n\u003cp\u003eLASSO logistic regression was used to select variables by shrinking regression coefficients with an L1 penalty, thereby reducing less informative features to zero. SVM-RFE was employed to iteratively eliminate features and identify the most relevant genes. RF analysis was performed to evaluate the importance of each variable based on decision tree models. Hub genes were determined by integrating the results of these three algorithms and visualized by Venn diagram. The expression patterns and diagnostic performance of hub genes were validated across different datasets. Receiver operating characteristic (ROC) curves and nomograms were generated using the \u0026lsquo;\u003cem\u003epROC\u003c/em\u003e\u0026rsquo;, \u0026lsquo;\u003cem\u003erms\u003c/em\u003e\u0026rsquo;, and \u0026lsquo;\u003cem\u003ermda\u003c/em\u003e\u0026rsquo; R packages to evaluate diagnostic efficacy of the hub genes found based on machine learning.\u003c/p\u003e\n\u003cp\u003eTo further explore the functional roles of hub genes, single-gene gene set enrichment analysis (GSEA) was performed using the \u0026lsquo;\u003cem\u003eclusterProfiler\u003c/em\u003e\u0026rsquo; package [16] based on the hallmark gene sets obtained from the MSigDB database[17]. Then, Immune cell infiltration was estimated using the \u0026lsquo;CIBERSORT \u0026lsquo;package[18] of R.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSingle-cell transcriptomic analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe processed single-cell RNA sequencing (scRNA-seq) dataset (GSE216009) was also obtained from the Gene Expression Omnibus (GEO) database. Given that the data were provided as a preprocessed \u0026lsquo;Seurat\u0026rsquo; object, no additional raw data preprocessing or quality control was performed. Downstream analyses were conducted using the \u0026lsquo;\u003cem\u003eSeurat\u003c/em\u003e\u0026rsquo;\u0026nbsp;R package. The expression patterns of hub genes were evaluated across different cell populations. In addition, gene perturbation analysis was performed using\u0026nbsp;\u0026lsquo;\u003cem\u003escTenifoldKnk\u003c/em\u003e\u0026rsquo;\u0026nbsp;R package [19] to assess the impact of hub gene knockout.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAI-driven drug prediction and molecular docking\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIntegrated bulk RNA-seq datasets were used as input to screen for potential drugs capable of modulating disease-associated gene expression patterns. The DrugRefLector [20] framework was applied to identify candidate therapeutic compounds based on transcriptomic profiles.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMolecular docking analysis was performed between candidate compounds and hub gene-encoded protein to further evaluate drug\u0026ndash;target interactions. Protein structures were obtained from the RCSB Protein Data Bank (PDB) [21], and ligand structures were retrieved from the PubChem database [22]. Protein structures were preprocessed using PyMOL to remove water molecules and bound ligands. Molecular docking was conducted using AutoDock Vina to predict binding affinities and identify potential binding sites. Docking scores (kcal/mol) were calculated for each protein\u0026ndash;ligand interaction, and the conformation with the lowest binding energy was selected as the optimal binding mode. The docking results were visualized using PyMOL to illustrate binding interactions, including hydrogen bond formation between ligands and target proteins.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll statistical analyses were performed in R software (Version 4.5.3). Differences between two groups were assessed using Student\u0026rsquo;s t-test. \u0026nbsp; A two-tailed p-value \u0026lt; 0.05 was considered statistically significant.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe study design is presented in Figure 1A. First of all, we integrated five bulk RNA-seq datasets from independent studies to construct a unified transcriptomic cohort for following analyses. Prior to batch correction, principal component analysis (PCA) revealed strong dataset-driven clustering, with samples segregating primarily according to their study of origin rather than biological differences between sepsis and healthy individuals (Figure 1B). This indicates the presence of substantial batch effects across datasets. Thereby, batch correction was performed to mitigate these confounding effects. Following correction, PCA demonstrated a marked reduction in dataset-specific clustering, with samples from different cohorts becoming more intermixed (Figure 1C). Consistently, global gene expression distributions across samples became more comparable after batch correction, as shown by the aligned expression profiles across datasets (Figure 1D), enabling reliable downstream integrative analyses.\u003c/p\u003e\n\u003cp\u003ePCA (Figure 2A) revealed a significant separation between sepsis patients and healthy individuals along PC1 (p \u0026lt; 0.001), despite partial overlap between groups. A total of 1,845 DEGs were identified, including 599 upregulated and 1246 downregulated genes (Figure 2B). Distinct expression profiles between patients with sepsis and healthy individuals was further illustrated by heatmap (Figure 2C).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMultiple gene co-expression modules across the dataset were found by WGCNA (Figure 2D). Module\u0026ndash;trait relationship analysis revealed several modules significantly associated with disease status (Figure 2E). In particular, the brown module showed strong positive correlation with disease, while the magenta module exhibited a negative association (Figure 2E). Consistent with the module\u0026ndash;trait correlation analysis (Figure 2E), the brown and magenta modules also exhibited the highest gene significance values (Figure 2F), further confirming their strong association with disease status.\u003c/p\u003e\n\u003cp\u003eA total of 4,178 genes were identified within the magenta module, of which 1,216 genes overlapped with differentially expressed genes (DEGs), as shown in the Venn diagram (Figure 3A). KEGG pathway enrichment analysis revealed that these overlapping genes were predominantly enriched in pathways associated with the ribosome, human T-cell leukemia virus 1 infection, coronavirus disease (COVID-19), and cytokine\u0026ndash;cytokine receptor interaction (Figure 3B). GO enrichment analysis further demonstrated that these genes were involved in immune-related biological processes (BP), including lymphocyte differentiation, leukocyte cell\u0026ndash;cell adhesion, and regulation of T cell activation. In terms of cellular components (CC), these genes were mainly enriched in ribosome-associated structures. Moreover, molecular function (MF) analysis indicated enrichment in catalytic activity, structural constituent of ribosome, and immune receptor activity (Figure 3C).\u003c/p\u003e\n\u003cp\u003eSimilarly, 1,268 genes were identified in the brown module, among which 440 genes overlapped with DEGs (Figure S1A). KEGG pathway analysis indicated that these genes were primarily enriched in cytokine\u0026ndash;cytokine receptor interaction and the MAPK signaling pathway (Figure S1B). GO enrichment analysis further revealed their significant enrichment in immune-related biological processes, such as regulation of immune effector processes, inflammatory response, and leukocyte-mediated immunity. In addition, molecular function analysis highlighted enrichment in immune receptor activity (Figure S1C).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTaken together, these findings suggest that the magenta module integrates both immune and translational processes, whereas the brown module predominantly represents canonical inflammatory signaling, highlighting the critical roles of translational dysregulation and inflammation in sepsis.\u003c/p\u003e\n\u003cp\u003eThe 1,216 overlapping genes (Figure 3A) were further analyzed using a protein\u0026ndash;protein interaction (PPI) network, revealing a densely interconnected network architecture (Figure 4A). Based on node connectivity, the top 10 hub genes were identified according to their degree centrality within the network (Figure 4B). To further assess candidate biomarkers, these hub genes were subjected to multiple machine learning approaches, including LASSO regression, SVM-RFE, and random forest analysis (Figure 4C\u0026ndash;H). Integrative analysis of these three algorithms identified five overlapping key genes\u0026mdash;RPL5, RPS23, RPS27A, RPS19, and RPS3A (Figure 4I). Notably, all identified genes are ribosome-associated proteins, further supporting the involvement of translational regulation in sepsis pathogenesis. A diagnostic nomogram based on these five genes was subsequently constructed to distinguish sepsis patients from healthy controls (Figure 4J). ROC curves demonstrated robust diagnostic performance of these 5 genes, among which RPL5 exhibited the highest predictive accuracy, with an area under the curve (AUC) of 0.859 (Figure 4K).\u003c/p\u003e\n\u003cp\u003eSimilarly, the 440 overlapping genes derived from the brown module (Figure S1A) were subjected to PPI network analysis, revealing a complex interaction network (Figure S2A). The top 10 hub genes were identified based on node connectivity (Figure S2B) and subsequently analyzed using LASSO regression, SVM-RFE, and random forest approaches (Figure S2C\u0026ndash;H). Integrative analysis identified seven overlapping key genes\u0026mdash;JAK2, FCER1G, IL1R1, JAK3, MAPK14, VSIG4, and CSF3R (Figure S2I), which are predominantly involved in classical inflammatory signaling pathways. A diagnostic nomogram based on these genes was constructed (Figure S2J), and ROC analysis demonstrated strong predictive performance of these key genes, with IL1R1 showing the highest diagnostic accuracy (AUC = 0.855) (Figure S2K).\u003c/p\u003e\n\u003cp\u003eCollectively, these findings further support a dual regulatory framework in sepsis, in which ribosome-associated genes (e.g., RPL5) represent the translational arm, whereas immune signaling genes (e.g., IL1R1) reflect the inflammatory arm, together contributing to the pathogenesis of sepsis.\u003c/p\u003e\n\u003cp\u003eGSEA was applied to further elucidate the molecular role of RPL5. Samples with high RPL5 expression were significantly enriched in multiple immune-related pathways, including cytokine\u0026ndash;cytokine receptor interaction and T cell\u0026ndash;associated signaling pathways (Figure 5A). Samples with low RPL5 expression were enriched in complement and coagulation cascades, as well as other immune-related pathways (Figure 5B). CIBERSORT analysis was conducted to further characterize the immune microenvironment associated with RPL5, and the results showed that RPL5 expression was significantly correlated with the proportions of several immune cell populations, including CD4 memory T cells (resting and activated), regulatory T cells (Tregs), and monocytes. Notably, RPL5 expression was negatively correlated with macrophage subsets (M0, M1, and M2) (Figure 5C), indicating that reduced RPL5 expression may be associated with enhanced macrophage-driven inflammatory responses and impaired adaptive immune regulation.\u003c/p\u003e\n\u003cp\u003eSimilarly, GSEA was performed to investigate the molecular role of IL1R1. It was observed that samples with high IL1R1 expression were significantly enriched in classical inflammatory signaling pathways, including MAPK signaling and Toll-like receptor signaling pathways (Figure S3A). In contrast, samples with low IL1R1 expression showed enrichment in pathways such as allograft rejection and asthma (Figure S3B). Consistent with these findings, CIBERSORT analysis revealed that IL1R1 expression was significantly associated with multiple immune cell populations, including memory B cells, CD8 T cells, monocytes, macrophages (M0 and M2), and neutrophils (Figure S3C).\u003c/p\u003e\n\u003cp\u003eSingle-cell RNA sequencing analysis was performed using the GSE216009 dataset to further validate the role of RPL5 at single-cell resolution, which included 12 healthy individuals and 52 sepsis patients. UMAP visualization identified 24 distinct cell types across all samples (Figure 6A), with neutrophils and T cells representing the most abundant populations. The expression pattern of RPL5 was subsequently examined, revealing that it is broadly expressed across multiple immune cell types, particularly in T cells, monocytes, plasmacytoid dendritic cells (pDCs), and hematopoietic stem and progenitor cells (HSPCs) (Figure 6B\u0026ndash;C), suggesting a potential role in coordinating immune cell function. To further understand the functional impact of RPL5, in silico knockout analysis was performed. Approximately 0.5% of genes were significantly differentially expressed following RPL5 knockout (Figure 6D). The top 20 DEGs included ANXA3, CEACAM8, STOM, CD24, and UGCG (Figure 6E). Functional enrichment analysis demonstrated that these DEGs were significantly involved in immune-related biological processes, such as defense response to bacterium and cell killing (Figure 6F). Consistently, KEGG pathway analysis revealed enrichment in immune-associated pathways, particularly the IL-17 signaling pathway (Figure 6G), highlighting a potential role for RPL5 in modulating host immune responses during sepsis.\u003c/p\u003e\n\u003cp\u003eLikewise, to assess the functional significance of IL1R1, in silico knockout analysis was conducted. Approximately 0.5% of genes were significantly differentially expressed following IL1R1 knockout (Figure S4A). The top 20 DEGs overlapped substantially with those observed in RPL5 knockout, including ANXA3, CEACAM8, CD24, STOM, and UGCG (Figure S4B). Functional enrichment analysis showed that these DEGs were significantly involved in immune-related biological processes, including defense response to bacterium, defense response to Gram-negative bacterium, and antimicrobial humoral response (Figure S4C). KEGG pathway analysis found their enrichment in cobalamin transport and metabolism (Figure S4D), suggesting additional metabolic regulation associated with IL1R1 signaling.\u003c/p\u003e\n\u003cp\u003eWe applied an AI-driven drug screening framework (DrugRefLector) using an integrated bulk transcriptomic profile derived from five microarray datasets. This analysis identified ten candidate compounds capable of reversing sepsis-associated transcriptional signatures (Figure 7A), among which BRD-K39768328 (Figure 7B) and BRD-K06569345 (Figure 7C) emerged as the top-ranked candidates. To assess whether RPL5 could serve as a direct target of these compounds, molecular docking analysis was performed. The results suggested that BRD-K39768328 binds to the C1 cavity pocket of RPL5 with a binding affinity of \u0026minus;7.3 kcal/mol (Figure 7D), while BRD-K06569345 binds to the C4 cavity pocket with a binding affinity of \u0026minus;7.5 kcal/mol (Figure 7E), indicating stable and favorable interactions.\u003c/p\u003e\n\u003cp\u003eMolecular docking analysis also provided the evidence that BRD-K39768328 binds to the C1 cavity pocket of IL1R1 with a binding affinity of \u0026minus;7.2 kcal/mol (Figure S5A), whereas BRD-K06569345 binds to the C2 cavity pocket with an even stronger binding affinity of \u0026minus;8.3 kcal/mol (Figure S5B), suggesting high binding stability. Notably, both compounds exhibited binding affinity toward RPL5 and IL1R1, indicating a potential dual-target therapeutic mechanism.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we performed an integrative multi-omics analysis combining bulk transcriptomics, single-cell RNA sequencing, and machine learning approaches to uncover key regulatory mechanisms in sepsis. Our results identified RPL5 and IL1R1 as two central hub genes, representing the translational regulation and inflammatory signaling axis, respectively. These findings suggest that simultaneous targeting of these two pathways may represent a promising therapeutic strategy for sepsis. Therefore, we applied an AI-driven drug screening framework and identified two candidate compounds, BRD-K39768328 and BRD-K06569345. Molecular docking analysis demonstrated that both compounds exhibit stable binding affinities toward RPL5 and IL1R1, supporting their potential as dual-target therapeutic agents.\u003c/p\u003e \u003cp\u003eAccumulating evidence suggests that ribosomal proteins are not only structural components of the translational machinery but also play critical roles in immune regulation. For instance, ribosomal protein S3 (RPS3) serves as a critical regulator in the NF-κB signaling pathway, playing an essential role in determining promoter selectivity and transcriptional specificity of NF-κB [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Another study has showed that silencing RPS3 can mitigate cigarette smoke-induced acute lung injury by inhibiting NF-κB activity [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Importantly, RPS3 has been reported to participate in the inflammatory response in sepsis-associated acute kidney injury (S-AKI) through activation of the NF-κB signaling pathway [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. In aligning with these findings, we also observed that ribosomal protein S3A (RPS3A) was a key gene within the magenta module and, together with other ribosome-associated genes, exhibited strong diagnostic value in distinguishing patients with sepsis from healthy individuals. Beyond RPS3, growing evidence indicates that other ribosome-associated genes are frequently identified as key regulators in sepsis-related conditions. Previous studies have demonstrated that several ribosomal genes serve as central hub genes in sepsis-related myopathy, suggesting that ribosome-associated processes may be broadly involved in immune dysregulation during sepsis [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBuilding upon these evidence, our study extends the role of ribosomal proteins to ribosomal protein L5 (RPL5), a ribosomal protein involved in ribosome biogenesis and translational control, which has been largely unexplored in the context of sepsis. In this study, RPL5 was found to be significantly downregulated in sepsis and was associated with ribosome-related pathways and immune processes. Its downregulation may impair ribosome function and disrupt translational capacity. Notably, we found that RPL5 was broadly expressed across multiple immune cell types, including T cells and monocytes, implying its potential role in maintaining immune cell function. Single-cell transcriptomic analysis further supported these findings by revealing cell-type-specific expression patterns of RPL5 and its involvement in immune cell populations. In silico gene perturbation analysis demonstrated that RPL5 knockout affected a functionally relevant subset of genes, particularly those involved in immune-related pathways such as IL-17 signaling, indicating a targeted regulatory role rather than a global transcriptional effect. Unlike RPS3, which has been shown to actively promote NF-κB signaling, our results indicate that RPL5 is significantly downregulated in sepsis, implying a distinct mechanism. We hypothesize that reduced RPL5 expression may impair ribosomal integrity and translational capacity, thereby contributing to immune dysregulation. Collectively, these findings highlight a broader role of ribosomal proteins in immune regulation and support the notion that dysregulation of ribosome-associated processes may represent an important mechanism underlying sepsis pathogenesis which warrant further investigation.\u003c/p\u003e \u003cp\u003eInterleukin-1 receptor type 1 (IL1R1) is a critical cytokine receptor belonging to the IL-1 receptor family, primarily acting as a high-affinity receptor for proinflammatory cytokines IL-1α and IL-1β. It mediates immune and inflammatory responses, including neutrophil recruitment and NF-\u0026#120581;B activation [\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Consistent with previous studies, our results demonstrated that IL1R1 is significantly upregulated in sepsis and is predominantly expressed in inflammatory immune cell populations, including monocytes, macrophages, and neutrophils. These findings support the role of IL1R1 as a key driver of immune dysregulation and cytokine storm in sepsis. We also found that IL1R1 exhibited strong diagnostic performance in distinguishing patients with sepsis from healthy individuals, indicating that it may serve not only as a potential therapeutic target but also as a clinically relevant diagnostic biomarker.\u003c/p\u003e \u003cp\u003eFrom a translational perspective, our work is the first study that identified plausible therapeutic compounds using an AI-driven drug prediction framework in the context of sepsis. Notably, BRD-K39768328 and BRD-K06569345 showed favorable binding affinities with both RPL5 and IL1R1, highlighting their potential to simultaneously modulate ribosome-associated and inflammatory pathways. This dual-target strategy may represent a promising direction for future therapeutic development in sepsis.\u003c/p\u003e \u003cp\u003eHowever, it must be noted that several limitations should be acknowledged. First, our study is primarily based on publicly available datasets and computational analyses, lacking experimental validation. Second, the mechanistic interactions between IL1R1 and RPL5 require further investigation in vitro and in vivo. Future studies are needed to validate these findings and explore their clinical applicability.\u003c/p\u003e \u003cp\u003eIn conclusion, our study identifies RPL5 and IL1R1 as reliable diagnostic biomarkers and therapeutic targets in sepsis. In addition, BRD-K39768328 and BRD-K06569345 were identified as promising candidate compounds with the potential to reverse sepsis-associated gene expression patterns, warranting further validation in preclinical models. These findings provide new insights into the pathogenesis of sepsis and offer potential avenues for the development of targeted therapeutic strategies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDeclaration of competing interests:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors state that they have no conflicts of interest, as per the journal\u0026apos;s definition.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical Approval and Consent to participate:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was conducted based entirely on publicly available datasets retrieved from the Gene Expression Omnibus (GEO) database. All data used in this study were previously published and approved by the respective institutional review boards. Therefore, no additional ethical approval or informed consent was required for this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of supporting data:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/). All code used in this study is available upon request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eM.L. conceived and designed the study. M.L. conducted statistical analyses and data visualization. M.L. wrote the original draft of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to thank the contributors of the Gene Expression Omnibus (GEO) database for providing publicly available datasets, which made this study possible.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSinger, M., et al., \u003cem\u003eThe Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3).\u003c/em\u003e Jama, 2016. \u003cstrong\u003e315\u003c/strong\u003e(8): p. 801\u0026ndash;10.\u003c/li\u003e\n\u003cli\u003eRudd, K.E., et al., \u003cem\u003eGlobal, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study.\u003c/em\u003e Lancet, 2020. \u003cstrong\u003e395\u003c/strong\u003e(10219): p. 200\u0026ndash;211.\u003c/li\u003e\n\u003cli\u003eLi, M., et al., \u003cem\u003eHigh-density lipoprotein: a biomarker and therapeutic target in sepsis.\u003c/em\u003e Crit Care, 2025. \u003cstrong\u003e29\u003c/strong\u003e(1): p. 453.\u003c/li\u003e\n\u003cli\u003eWiersinga, W.J., et al., \u003cem\u003eHost innate immune responses to sepsis.\u003c/em\u003e Virulence, 2014. \u003cstrong\u003e5\u003c/strong\u003e(1): p. 36\u0026ndash;44.\u003c/li\u003e\n\u003cli\u003eWillmann, K. and L.F. Moita, \u003cem\u003ePhysiologic disruption and metabolic reprogramming in infection and sepsis.\u003c/em\u003e Cell Metab, 2024. \u003cstrong\u003e36\u003c/strong\u003e(5): p. 927\u0026ndash;946.\u003c/li\u003e\n\u003cli\u003eCao, M., G. Wang, and J. Xie, \u003cem\u003eImmune dysregulation in sepsis: experiences, lessons and perspectives.\u003c/em\u003e Cell Death Discov, 2023. \u003cstrong\u003e9\u003c/strong\u003e(1): p. 465.\u003c/li\u003e\n\u003cli\u003eZhang, Y.Y. and B.T. Ning, \u003cem\u003eSignaling pathways and intervention therapies in sepsis.\u003c/em\u003e Signal Transduct Target Ther, 2021. \u003cstrong\u003e6\u003c/strong\u003e(1): p. 407.\u003c/li\u003e\n\u003cli\u003eWeber, A., P. Wasiliew, and M. Kracht, \u003cem\u003eInterleukin-1 (IL-1) pathway.\u003c/em\u003e Sci Signal, 2010. \u003cstrong\u003e3\u003c/strong\u003e(105): p. cm1.\u003c/li\u003e\n\u003cli\u003eJiao, L., et al., \u003cem\u003eRibosome biogenesis in disease: new players and therapeutic targets.\u003c/em\u003e Signal Transduct Target Ther, 2023. \u003cstrong\u003e8\u003c/strong\u003e(1): p. 15.\u003c/li\u003e\n\u003cli\u003eSundaramoorthy, E., et al., \u003cem\u003eRibosome quality control activity potentiates vaccinia virus protein synthesis during infection.\u003c/em\u003e J Cell Sci, 2021. \u003cstrong\u003e134\u003c/strong\u003e(8).\u003c/li\u003e\n\u003cli\u003eWang, H., et al., \u003cem\u003ePreliminary screening of new biomarkers for sepsis using bioinformatics and experimental validation.\u003c/em\u003e PLoS One, 2025. \u003cstrong\u003e20\u003c/strong\u003e(1): p. e0317608.\u003c/li\u003e\n\u003cli\u003eDavis, S. and P.S. Meltzer, \u003cem\u003eGEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor.\u003c/em\u003e Bioinformatics, 2007. \u003cstrong\u003e23\u003c/strong\u003e(14): p. 1846\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eRitchie, M.E., et al., \u003cem\u003elimma powers differential expression analyses for RNA-sequencing and microarray studies.\u003c/em\u003e Nucleic Acids Res, 2015. \u003cstrong\u003e43\u003c/strong\u003e(7): p. e47.\u003c/li\u003e\n\u003cli\u003eGu, Z., \u003cem\u003eComplex heatmap visualization.\u003c/em\u003e Imeta, 2022. \u003cstrong\u003e1\u003c/strong\u003e(3): p. e43.\u003c/li\u003e\n\u003cli\u003eLangfelder, P. and S. Horvath, \u003cem\u003eWGCNA: an R package for weighted correlation network analysis.\u003c/em\u003e BMC Bioinformatics, 2008. \u003cstrong\u003e9\u003c/strong\u003e: p. 559.\u003c/li\u003e\n\u003cli\u003eYu, G., et al., \u003cem\u003eclusterProfiler: an R package for comparing biological themes among gene clusters.\u003c/em\u003e Omics, 2012. \u003cstrong\u003e16\u003c/strong\u003e(5): p. 284\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eLiberzon, A., et al., \u003cem\u003eThe Molecular Signatures Database (MSigDB) hallmark gene set collection.\u003c/em\u003e Cell Syst, 2015. \u003cstrong\u003e1\u003c/strong\u003e(6): p. 417\u0026ndash;425.\u003c/li\u003e\n\u003cli\u003eChen, B., et al., \u003cem\u003eProfiling Tumor Infiltrating Immune Cells with CIBERSORT.\u003c/em\u003e Methods Mol Biol, 2018. \u003cstrong\u003e1711\u003c/strong\u003e: p. 243\u0026ndash;259.\u003c/li\u003e\n\u003cli\u003eOsorio, D., et al., \u003cem\u003escTenifoldKnk: An efficient virtual knockout tool for gene function predictions via single-cell gene regulatory network perturbation.\u003c/em\u003e Patterns (N Y), 2022. \u003cstrong\u003e3\u003c/strong\u003e(3): p. 100434.\u003c/li\u003e\n\u003cli\u003eDeMeo, B., et al., \u003cem\u003eActive learning framework leveraging transcriptomics identifies modulators of disease phenotypes.\u003c/em\u003e Science, 2025. \u003cstrong\u003e390\u003c/strong\u003e(6776): p. eadi8577.\u003c/li\u003e\n\u003cli\u003eBerman, H.M., et al., \u003cem\u003eThe Protein Data Bank.\u003c/em\u003e Nucleic Acids Res, 2000. \u003cstrong\u003e28\u003c/strong\u003e(1): p. 235\u0026ndash;42.\u003c/li\u003e\n\u003cli\u003eKim, S., \u003cem\u003eGetting the most out of PubChem for virtual screening.\u003c/em\u003e Expert Opin Drug Discov, 2016. \u003cstrong\u003e11\u003c/strong\u003e(9): p. 843\u0026ndash;55.\u003c/li\u003e\n\u003cli\u003eWan, F., et al., \u003cem\u003eIKK\u0026beta; phosphorylation regulates RPS3 nuclear translocation and NF-\u0026kappa;B function during infection with Escherichia coli strain O157:H7.\u003c/em\u003e Nat Immunol, 2011. \u003cstrong\u003e12\u003c/strong\u003e(4): p. 335\u0026ndash;43.\u003c/li\u003e\n\u003cli\u003eDong, J., et al., \u003cem\u003eRibosomal Protein S3 Gene Silencing Protects Against Cigarette Smoke-Induced Acute Lung Injury.\u003c/em\u003e Mol Ther Nucleic Acids, 2018. \u003cstrong\u003e12\u003c/strong\u003e: p. 370\u0026ndash;380.\u003c/li\u003e\n\u003cli\u003eZhang, X., et al., \u003cem\u003eRPS3 Aggravates Sepsis-Induced Acute Kidney Injury Through Activating NF-kappaB Mediated Renal Inflammatory Responses.\u003c/em\u003e Physiol Res, 2026. \u003cstrong\u003e75\u003c/strong\u003e(1): p. 63\u0026ndash;77.\u003c/li\u003e\n\u003cli\u003eWang, J., K. Han, and J. Lu, \u003cem\u003eScreening of hub genes for sepsis-induced myopathy by weighted gene co-expression network analysis and protein-protein interaction network construction.\u003c/em\u003e BMC Musculoskelet Disord, 2024. \u003cstrong\u003e25\u003c/strong\u003e(1): p. 834.\u003c/li\u003e\n\u003cli\u003eRalph, B.A., et al., \u003cem\u003eThe IL-1 Receptor Is Required to Maintain Neutrophil Viability and Function During Aspergillus fumigatus Airway Infection.\u003c/em\u003e Front Immunol, 2021. \u003cstrong\u003e12\u003c/strong\u003e: p. 675294.\u003c/li\u003e\n\u003cli\u003eGuo, Q., et al., \u003cem\u003eNF-\u0026kappa;B in biology and targeted therapy: new insights and translational implications.\u003c/em\u003e Signal Transduct Target Ther, 2024. \u003cstrong\u003e9\u003c/strong\u003e(1): p. 53.\u003c/li\u003e\n\u003cli\u003eGehrke, N., et al., \u003cem\u003eHepatocyte-specific deletion of IL1-RI attenuates liver injury by blocking IL-1 driven autoinflammation.\u003c/em\u003e J Hepatol, 2018. \u003cstrong\u003e68\u003c/strong\u003e(5): p. 986\u0026ndash;995.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Nanjing Drum Tower Hospital","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Sepsis, RPL5, IL1R1, Multi-omics integration, Drug prediction","lastPublishedDoi":"10.21203/rs.3.rs-9624193/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9624193/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eSepsis is a life-threatening condition characterized by dysregulated immune responses to infection. While advances in critical care have improved, sepsis-related mortality remains high, underscoring the urgent need for further research into therapeutic targets and potential candidate agents.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eWe performed an integrative multi-omics analysis combining five bulk transcriptomic datasets and single-cell RNA sequencing data. Differential expression analysis and WGCNA were used to identify disease-associated modules. Machine learning algorithms, including LASSO, SVM-RFE, and random forest, were applied to screen key hub genes. Functional enrichment, immune infiltration analysis, and in silico gene perturbation were conducted to explore biological roles of these hub genes. An AI-driven drug prediction framework together with molecular docking were used to identify potential therapeutic compounds.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eWe identified 1,845 differentially expressed genes and sepsis-associated modules, among which the magenta and brown modules found by WGCNA analysis showed the strongest correlation with sepsis. Integrative analysis identified RPL5 and IL1R1 as key hub genes, with RPL5 demonstrating the highest diagnostic performance (AUC = 0.859). Functional analyses revealed that RPL5 is associated with ribosomal pathways and immune regulation, whereas IL1R1 is linked to inflammatory signaling. Single-cell analysis showed that RPL5 is broadly expressed across multiple immune cell types. In silico knockout indicated that RPL5 regulates immune-related pathways, including IL-17 signaling. Drug prediction identified BRD-K39768328 and BRD-K06569345 as candidate compounds with favorable binding affinities to both RPL5 and IL1R1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eThis study identified RPL5 and IL1R1 as potential diagnostic and therapeutic targets and proposed BRD-K39768328 and BRD-K06569345 as candidate compounds for sepsis, offering novel insights into sepsis pathogenesis and potential therapeutic interventions.\u003c/p\u003e","manuscriptTitle":"Integrative multi-omics analysis identifies RPL5 and IL1R1 as potential diagnostic and therapeutic targets of sepsis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-07 08:50:10","doi":"10.21203/rs.3.rs-9624193/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d08a4104-cb57-4d11-b24f-086309e2405b","owner":[],"postedDate":"May 7th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":67596147,"name":"Critical Care \u0026 Emergency Medicine"}],"tags":[],"updatedAt":"2026-05-07T08:50:10+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-07 08:50:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9624193","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9624193","identity":"rs-9624193","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}