Network Medicine Approach to Viral Infections: Unraveling Universal and Virus-Specific Immune Pathways

Network Medicine Approach to Viral Infections: Unraveling Universal and Virus-Specific Immune Pathways | 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 Article Network Medicine Approach to Viral Infections: Unraveling Universal and Virus-Specific Immune Pathways Vanessa Escolano Maso, Ana Paula Barbosa, Viviane Schuch, Thomas Hagan, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5240411/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 13 You are reading this latest preprint version Abstract Viral infections continue to pose significant global health challenges, with pathogens like Dengue Virus (DENV), Chikungunya Virus (CHIKV), Yellow Fever Virus (YFV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU) contributing to widespread morbidity and mortality. A comprehensive understanding of the host immune response to these infections is crucial for developing more effective diagnostics, treatments, and vaccines. In this study, we conducted a meta-analysis of transcriptomic datasets from whole blood and peripheral blood mononuclear cell (PBMC) samples of individuals infected with DENV, CHIKV, YFV, HIV, and FLU. Utilizing a systems biology approach, we identified both shared and virus-specific molecular mechanisms underlying immune responses to these infections. Several key genes, including OAS2, MX1, and IFI44L, were consistently up-regulated across multiple viral infections, highlighting their roles in antiviral defense, immune modulation, and cellular stress responses. Additionally, virus-specific genes such as TNFSF10 and NT5C3A were found to be uniquely up-regulated in arboviruses, underscoring the distinct immune pathways activated by different viral families. Our co-expression network analysis revealed coordinated gene expression patterns across different viral infections, offering insights into conserved molecular networks that could inform the development of broad-spectrum antiviral therapies. Despite challenges related to biological heterogeneity and variations in study design, this meta-analysis provides a foundation for future research aimed at unraveling the complex immune mechanisms that govern host-pathogen interactions. Biological sciences/Computational biology and bioinformatics Biological sciences/Immunology/Immunogenetics Viral infections Immune response Co-expression networks Transcriptomics Meta-analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Viral infections remain a significant global health challenge, contributing to widespread morbidity and mortality across the world 1 . Pathogens such as Dengue Virus (DENV), Chikungunya Virus (CHIKV), Yellow Fever Virus (YFV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU) infect millions of people annually, causing a wide range of clinical outcomes and posing an ongoing threat to public health systems 1 . These viruses have distinct epidemiological characteristics and transmission routes, with symptoms that can vary from mild to life-threatening. For instance, DENV can lead to severe hemorrhagic manifestations 2 , while CHIKV often causes chronic arthralgia 3 , and YFV is associated with potentially fatal liver failure 4 . Understanding the host immune response to these infections is essential for the development of more effective diagnostics, treatments, and vaccines. However, the molecular mechanisms that drive these immune responses, particularly the differentiation between shared and virus-specific pathways, are not fully understood, making this a critical area of research for improving the management of viral diseases. Systems biology and network medicine approaches have transformed our understanding of host-pathogen interactions by emphasizing the interconnectedness of molecular processes during viral infections 5 – 7 . Rather than focusing on individual genes, these approaches map the complex interactions between genes, proteins, and cellular pathways, offering a more holistic view of how the immune system responds to viral threats. Transcriptomic studies, which capture genome-wide gene expression profiles, are essential tools for constructing these interaction networks 8 . By identifying differentially expressed genes (DEGs), researchers can pinpoint key nodes and pathways that play pivotal roles in immune responses, disease progression, and recovery. Analysis of whole blood and peripheral blood mononuclear cells (PBMCs) during viral infections has revealed crucial genes and pathways involved in antiviral defense, inflammation, and immune modulation 9 – 11 . However, while transcriptomic data is valuable for generating these molecular networks, individual studies often lack the statistical power to provide a comprehensive understanding of the system 12 . Thus, integrating data across multiple studies and viral infections is crucial to uncover both shared immune mechanisms and virus-specific responses, ultimately revealing conserved molecular patterns that can inform broad-spectrum antiviral therapies and identify unique pathways for virus-specific treatments. Despite the vast availability of transcriptomic datasets, comparing gene expression patterns across different viruses using a systems biology approach remains challenging. Variations in study design, sample sizes, and the use of different technologies, such as microarray and RNA-seq, can introduce biases that complicate the construction of accurate molecular networks. Moreover, biological heterogeneity among infected individuals further complicates the interpretation of network structures, as variations in age, genetic background, or comorbidities can obscure consistent molecular signatures. To our knowledge, no study has systematically employed a transcriptomic meta-analysis across multiple viral infections to identify consistent molecular signatures and network structures. A network medicine approach, integrating data from different viruses, holds the potential to reveal both common and virus-specific molecular networks, offering deeper insights into immune responses and uncovering potential therapeutic targets. In this study, we applied a systems biology approach to perform a meta-analysis of transcriptomic datasets from whole blood and PBMC samples of individuals acutely infected with DENV, CHIKV, YFV, HIV, and FLU. Our primary objective was to construct robust molecular networks that capture the immune responses to these viral infections by integrating data from multiple studies and platforms. By leveraging these networks, we sought to uncover both shared and virus-specific molecular mechanisms and pathways. Through pathway enrichment analyses, we further elucidated the biological processes and immune responses activated by each virus. Our findings contribute to a deeper understanding of the universal and virus-specific immune pathways involved in viral infections, offering potential avenues for therapeutic interventions, and enhancing our knowledge of host-pathogen interactions in the context of multiple viral threats. Material and Methods Study Selection and Data Collection The studies were identified through systematic searches in the Gene Expression Omnibus (GEO) database 13 . The following six inclusion criteria were applied: 1) samples from humans in the acute phase of infection by one of the five viruses—DENV, YFV, CHIKV, FLU, or HIV—as well as samples from healthy individuals and/or those in the convalescent phase of viral infection; 2) sample donors must not have received any medicinal treatments; 3) samples must be obtained from whole blood or peripheral blood mononuclear cells (PBMCs); 4) transcriptomic studies must have been conducted using microarray or bulk RNA-seq technologies; 5) sequencing platforms must belong to either Illumina or Affymetrix; 6) a minimum of 20 samples must be available after excluding outliers. The minimum sample size was determined based on the number required to construct robust co-expression networks, meaning networks whose topology remains stable as additional samples are include 14 . Studies meeting all six inclusion criteria were incorporated into our meta-analysis. We selected transcriptomic datasets from both microarray and RNA-seq platforms for five viral infections or vaccination with attenuated virus: DENV, YFV, CHIKV, Influenza Virus (FLU), and HIV. For DENV, six microarray studies (GSE13052, GSE25001, GSE28405, GSE28988, GSE29891, GSE51808) and one RNA-seq study (GSE157240) were included, comprising a total of 532 infected and 212 non-infected control samples. The YFV dataset comprises of PBMC from 14 convalescent and 41 acutely infected individuals (GSE243442) using RNA-seq. For CHIKV, two RNA-seq studies (GSE99992, PRJNA507472) contributed a total of 81 infected and 60 control samples. FLU datasets included two microarray studies (GSE29366, GSE34205) and one RNA-seq study (GSE157240), totaling 154 samples (112 infected, 42 control). For HIV, two microarray datasets (GSE29429 using the GPL10558 platform and GSE29429 using the GPL6947 platform) were included, providing 56 infected and 53 control samples. Data processing and Normalization Raw microarray gene expression data were downloaded from GEO ( https://www.ncbi.nlm.nih.gov/geo/ ) using the GEOquery package 15 . Data were imported into RStudio using the ReadAffy function from the affy package for Affymetrix platforms and the read.ilmn function from the limma package for Illumina platforms. Quality control was conducted using the arrayQualityMetrics function from the arrayQualityMetrics package 16 . Background correction, quantile normalization, and log2 transformation were performed using the neqc function from the limma package for Illumina data and the justRMA function from affy for Affymetrix data. Quality control of normalized data included identifying and removing outlier samples using the arrayQualityMetrics package. Normalized data were further scrutinized for outliers, defined as samples flagged by at least two of the following tests: i) mean absolute difference of M-values (log-ratios) between arrays, ii) the Kolmogorov-Smirnov statistic (Ka) comparing each array’s signal distribution to pooled data, and iii) Hoeffding’s statistic (Da) on the joint distribution of A (average) and M values. Probes were annotated at the gene level according to the human genome version GRCh38.p13, Ensembl 104 17 , using the reannotator_microarray_probes pipeline, which aligns probe nucleotide sequences to the reference genome. Summarization retained the probe with the highest mean expression for each gene. Phenotypic outliers were identified using the Molecular Degree of Perturbation (MDP) package 18 , which evaluates the expression of the top 25% most perturbed genes. Identified outliers were excluded from subsequent analyses. Raw RNA-seq data were obtained from the Sequence Read Archive (SRA) using prefetch from the SRA Toolkit. Metadata for each study were retrieved using the SRA Run Selector. The SRA files were converted to fastq format with fasterq-dump from the SRA Toolkit. Quality control, adapter removal, and filtering of low-quality reads were performed using the fastp package 19 . For paired-end reads, the parameters --detect_adapter_for_pe, --cut_front, --cut_right, --qualified_quality_phred 20, --correction, and --overrepresentation_analysis were used. The same parameters were applied for single-end reads, except --detect_adapter_for_pe and --correction, which are specific to paired-end data. Reads were aligned to the human genome (GRCh38.p13, Ensembl release 104) using the STAR package 20 . Genomic indices were constructed using the primary assembly FASTA file and the corresponding GTF annotation file with the --sjdbOverhang 100 parameter. Aligned reads were stored in BAM files and sorted by genomic coordinates. Read summarization was conducted using featureCounts. For paired-end data, only read pairs aligned to the genome were counted. Chimeric fragments, multi-mapping, and multi-overlapping reads were excluded. Summarization was performed at the gene level using Ensembl gene identifiers (ENSG), considering strandedness according to the methodology reported in each study. Genes with zero counts or low expression were removed from the count table. Low expression was defined as fewer than one count per million (CPM) in at least x samples, where x is the sample size of the smallest phenotypic group. Library size was calculated by summing the counts of retained genes. Library size was normalized using the Trimmed Mean of M-values (TMM) method implemented in the calcNormFactors function of the edgeR package 21 . The count table was normalized using the CPM method implemented by the cpm function from edgeR. Biological outliers in the RNA-seq data were identified using the MDP test, which assesses the expression of the 25% most perturbed genes. Identified outliers were excluded from subsequent analyses. Differential Gene Expression Meta-Analysis For individual microarray studies, differentially expressed genes (DEGs) were identified using the limma package in R 22 , which employs a moderated t-test. For bulk RNA-seq studies, differential expression analysis was performed using the edgeR package. DEG analysis was conducted using the quasi-likelihood F-test, recommended for bulk RNA-seq data with biological replicates. DEGs were identified in each study by comparing samples from the acute phase of infection with control group samples, which included healthy and/or convalescent individuals. Meta-analysis was performed using Fisher's method for combining p-values, implemented in the MetaVolcanoR package 23 . A meta-analysis was conducted for each acute viral infection studied: DENV, CHIKV, FLU, and HIV. However, meta-analysis for YFV was not possible due to the availability of only one human blood transcriptome study for this infection. The p-values obtained from the meta-analysis were adjusted for multiple testing using the Benjamini-Hochberg False Discovery Rate (FDR-BH) method. Log2 fold change (log2FC) was calculated as the median log2FC for each gene across studies of the same infection. Genes were considered consistently differentially expressed for infection if they had an FDR-BH 1. Disease-gene networks Two disease-gene networks were constructed: one for consistently up-regulated genes and another for consistently down-regulated genes. The networks were built using Gephi software. Community detection in each network was performed using the Louvain algorithm. For network visualization, nodes were colored according to their community, with each community represented by a distinct color. The ForceAtlas algorithm was applied to spatialize the networks. Over-representation analysis (ORA) was conducted separately for the lists of consistently up-regulated and down-regulated genes from each infection analyzed in this study: DENV, CHIKV, FLU, and HIV. For YFV, the list of DEGs was used instead, due to the availability of only one dataset. ORA was performed using the clusterProfiler package 24 , which applies a right-tailed Fisher’s exact test. Blood Transcription Modules (BTMs) were used as the reference gene set, as they represent groups of genes previously identified as expressed in blood samples 25 . Gene sets were considered enriched if they had an FDR-BH < 0.05. Coexpression consensus network construction Gene coexpression networks and their modules were constructed using the CEMiTool R package 14 with the normalized gene expression matrix after removing outlier samples. Pearson positive correlations were used to identify modules. In bulk RNA-seq studies, the dependence between the mean and variance of gene expression was assessed. When such dependence was detected, the normalized gene expression matrix was transformed using the Variance Stabilizing Transformation (VST) method. A consensus coexpression network was constructed for each acute viral infection by creating a list of edges from the coexpression network of each study, then comparing these lists across all studies for the same infection, retaining only the edges present in more than half of the studies. Networks were visualized using Gephi software with the ForceAtlas2 algorithm. Community detection was performed using the Louvain algorithm. Centrality measures for the nodes, including degree, betweenness centrality, and harmonic centrality, were calculated. The biological functions associated with each community in the consensus coexpression networks were identified through ORA functional enrichment analysis, implemented with clusterProfiler 24 , using BTMs as reference gene sets. Enriched BTMs were identified based on an FDR-BH < 0.05. The communities in the consensus coexpression networks for DENV, CHIKV, FLU, and HIV, as well as the coexpression modules for YFV, were compared to identify similar or unique communities among the viral infections. This comparison was performed using the GeneOverlap package 26 , which conducts a right-tailed Fisher’s exact test to assess the overlap between two gene sets, with each community treated as a gene set. Additionally, the Jaccard index was calculated using GeneOverlap to measure the similarity between community pairs. Results Meta-Analysis of Transcriptomic Datasets Across Different Viral Infections We analyzed transcriptomic datasets of blood and peripheral blood mononuclear cells (PBMC) from individuals acutely infected with five different viruses: Yellow Fever Virus (YFV), Chikungunya Virus (CHIKV), Dengue Virus (DENV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU). Our analysis aimed to identify consistent and robust signatures of viral infection and to reveal potential molecular mechanisms shared or specific to each infection. The number of studies and samples included in our analysis varied by virus. We included one study for YFV (41 infected and 14 control samples), two studies for CHIKV (80 infected and 60 control samples), seven studies for DENV (316 infected and 213 control samples), two studies for HIV (56 infected and 53 control samples), and three studies for FLU (112 infected and 42 control samples) (Fig. 1 A). Our analytical workflow consisted of several key steps (see methods). We started with the collection of raw transcriptomic data, followed by quality control and normalization of gene expression levels. We then employed the Molecular Degree of Perturbation (MDP) method 18 to detect outliers. Subsequent steps included the identification of differentially expressed genes (DEGs) and co-expression patterns, followed by functional analysis. This workflow ensured the robustness and reliability of our meta-analysis (Fig. 1 B). To address sample heterogeneity, we applied the MDP method to all transcriptomic datasets. MDP quantifies the perturbation score of samples using nonperturbed subjects (i.e., healthy controls) as a reference group. This score allows the detection of outlying samples and subgroups of diseased patients, which is crucial because individual heterogeneity can significantly impact gene expression analyses. Figure 1 C and D illustrate a single example of this improvement using the Dengue virus study (GSE28988). The MDP scores for control samples (light blue) and infected samples (black) revealed several potential outliers (shown in red). Identifying and removing these outliers was crucial, as they can significantly impact the accuracy of DEG analysis (Fig. 1 C). For instance, the removal of outliers in GSE28988 dataset unveiled the T cell receptor signaling pathway (WP69), including genes such as MAP4K1, ITK, CD247, CD3E, ICOS, and SKAP1, which were not detected when outliers were included (Fig. 1 D). This underscores the importance of addressing sample heterogeneity to uncover biologically relevant pathways and genes. Next, we conducted a comprehensive meta-analysis of transcriptomic datasets from blood and PBMC samples of individuals acutely infected with these five different viruses. This analysis aimed to identify consistent and robust signatures of viral infection and uncover potential molecular mechanisms that are either shared across or specific to different infections. The number of up-regulated and down-regulated DEGs identified in each study for the five viruses varied across different studies (Fig. 2 A). Notably, DENV and CHIKV had a higher number of up-regulated genes compared to the other viruses (Fig. 2 A). We then performed a meta-analysis of DEGs using Fisher's method (Fig. 2 B, Table S1 ). This analysis aggregated the DEGs from individual studies to identify meta-DEGs, representing consistently differentially expressed genes across multiple studies (Fig. 2 B). Again, CHIKV and DENV showed the highest number of meta-DEGs, underscoring the robustness of these signatures across different datasets (Fig. 2 B). Pathway enrichment analysis was conducted on the meta-DEGs using over-representation analysis with blood transcription modules (BTM) as gene sets (Fig. 2 C). This analysis revealed distinct pathway enrichment profiles for each virus, providing insights into the specific immune responses and biological processes activated during infection (Fig. 2 C). To check for shared and specific meta-DEGs between the different viruses, we created a network using up-regulated meta-DEGs (Fig. 2 D). This network highlights genes that are shared among all viruses and those specific to arboviruses (DENV, CHIKV, YFV). Genes such as SPATS2L, OASL, and HERC6, which are shared by multiple viruses, may play crucial roles in the common antiviral response, while other genes are unique to specific viral infections, suggesting virus-specific pathways and mechanisms (Table S1 ). For instance, TNFSF10, PLAC8, and NT5C3A are uniquely up-regulated in arboviruses like DENV, CHIKV, and YFV, indicating virus-specific pathways and mechanisms (Fig. 2 D). This distinction between shared and unique genes provides insight into both universal and virus-specific immune responses. To further explore the molecular mechanisms underlying different viral infections, we performed a consensus network analysis of co-expression modules across the transcriptomic datasets 27 . This analysis aimed to identify genes and pathways that are consistently co-expressed across different viruses, thereby highlighting potential shared and virus-specific mechanisms. The number of genes included in the consensus co-expression network for each virus varied significantly, with YFV having the highest number of genes (2,223) and DENV the lowest (397) (Fig. 3 A). The number of community modules identified within the consensus network also varied, with CHIKV having the highest number of modules (31) and YFV the lowest (8) (Fig. 3 B, Table S2 ). The DENV consensus co-expression network highlights 18 community modules (C0 to C17), each represented by a different color (Fig. 3 C). These modules represent distinct groups of genes that are co-expressed during DENV infection, indicating coordinated biological processes. We identified various biological processes and cell-type related genes within these community modules. For example, genes related to neutrophil activity (e.g., LTF, DEFA1, CTSG, AZU1, MPO) and plasmablast presence (e.g., CD38, XBP1, TNFRSF17, JCHAIN, MZB1, CXCR3, TOP2A) were found in specific modules. Other modules included genes related to platelet function (e.g., ITGA2B, PF4V1, ITGB3, CLEC5A, CLEC1B, CLEC4F, CLEC12A), T cell response (e.g., CD8A, GZMK, GZMA, IL7R, CCR7, CXCR5), inflammation and antiviral response (e.g., TLR7, TLR4, ISG15, OAS1, OAS2, OAS3, OASL, IFIT1, IFIT2, IFIT3, IFIT5, MX1, MX2, CXCL10, GBP1, GBP5, GBP6, GBP4, GBP3), and complement system involvement (e.g., C4BPA, C1QA, C1QB, C1QC, C3AR1, SERPING1) (Fig. 3 D, Table S2 ). To identify shared and virus-specific co-expression patterns, we analyzed the overlap of genes between community modules from different viruses. The overlap analysis revealed significant co-expression patterns across the viral infections, with notable overlaps highlighted. For example, the black highlighted overlaps included genes related to the antiviral response such as ATF3, CCL2, and CXCL10 (Fig. 4 B). The green highlighted overlaps included genes associated with plasma B cells and plasmablasts, such as CD27, CD38, and CXCR3 (Fig. 4 C). Additionally, the brown and blue highlighted overlaps revealed genes involved in hemoglobin, heme metabolism, erythropoiesis (e.g., AHSP, HBA1, HBB, SLC25A39), and platelet function (e.g., ALOX12, ITGA2B, GP6, GP9) respectively (Fig. 4 D). Discussion In this study, we identified several consistently up-regulated genes across different viral infections, such as OAS2, MX1, and IFI44L, which, although commonly associated with the interferon response 28 , highlight additional layers of complexity in antiviral defense mechanisms. These genes are part of broader pathways involved not only in direct antiviral activity but also in modulating immune responses, inflammation, and cellular stress. For example, OAS2 has roles beyond RNA degradation, influencing immune signaling cascades, while MX1 has been shown to interact with other host proteins to limit viral replication. These findings underscore the importance of these genes in coordinating a more nuanced response to viral infections and highlight their potential as biomarkers for disease severity or progression. Furthermore, the expression of genes like IFI44L suggests possible roles in modulating the immune environment, contributing to the differentiation between acute and chronic infection stages across multiple viral threats. This suggests that while interferon response is a common theme, these genes may also be involved in more specialized functions that tailor immune responses to specific viral challenges. While we found shared molecular signatures, several virus-specific pathways also emerged from our analysis. For instance, genes like TNFSF10 and NT5C3A were up-regulated in arboviruses such as DENV, CHIKV, and YFV, but not in HIV or FLU. This suggests that certain immune pathways are more relevant to particular viral families, possibly reflecting unique interactions between the virus and host immune system. Pathway enrichment analysis further supported this, with pathways related to immune regulation, such as cytokine signaling and T cell receptor signaling, being more pronounced in specific viral infections. These findings open opportunities for virus-specific therapeutic interventions, as targeting these unique pathways could provide more tailored treatments for diseases like DENV and CHIKV. Despite these robust findings, there were several challenges in our analysis. Variations in study design, differences in sample sizes, and the use of both microarray and RNA-seq technologies introduced potential biases. Furthermore, biological heterogeneity within patient populations, including factors such as age, genetic background, and co-morbidities, complicated the identification of consistent molecular signatures. Although the MDP method allowed us to address some of this heterogeneity by identifying and removing outliers, these complexities underscore the need for more standardized datasets and methodologies in transcriptomic meta-analyses. Additionally, while our co-expression network analysis uncovered significant shared and virus-specific pathways, the extent to which these findings can be generalized across other viral infections requires further investigation. Our study provides valuable insights into both shared and virus-specific immune responses to viral infections, contributing to the broader understanding of host-pathogen interactions. The identification of conserved molecular networks across different viruses offers potential targets for broad-spectrum antiviral therapies, while the virus-specific pathways suggest avenues for more personalized treatments. Moving forward, expanding the analysis to include additional viral infections and larger datasets will further strengthen the understanding of these immune mechanisms. Future research should also focus on integrating multi-omic data, such as proteomics and metabolomics, with transcriptomic data to provide a more comprehensive view of host responses to viral infections. Additionally, the development of standardized protocols for transcriptomic studies could help mitigate the technical biases observed in our analysis, ultimately enhancing the reproducibility and robustness of meta-analyses in systems biology. Declarations Disclosures The authors declare no conflict of interest. Funding This work was supported by the São Paulo State Research Foundation (FAPESP) (grant numbers: 2018/14933-2 to H.I.N; 2023/05005-2 and 2019/16418-0 to V.E.M.). Author Contribution Conceptualization: V.E.M., H.I.N.; Data curation: All authors; Formal analysis: V.E.M., A.P.B.N.O., V.S., D.S.L.; Funding acquisition: H.I.N.; Methodology: All authors; Writing – original draft: V.E.M., H.I.N.; Writing – review and editing: All authors; Supervision: T.H., H.I.N. Acknowledgement The authors would like to thank all the members of Computational Systems Biology Laboratory (CSBL) for their input. References Naghavi, M. et al. Global burden associated with 85 pathogens in 2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet Infectious Diseases 24, 868–895 (2024). Kularatne, S. A. & Dalugama, C. Dengue infection: Global importance, immunopathology and management. Clin Med (Lond) 22, 9–13 (2022). Schilte, C. et al. Chikungunya Virus-associated Long-term Arthralgia: A 36-month Prospective Longitudinal Study. PLOS Neglected Tropical Diseases 7, e2137 (2013). Kallas, E. G. et al. Predictors of mortality in patients with yellow fever: an observational cohort study. The Lancet Infectious Diseases 19, 750–758 (2019). Farooq, Q. ul A., Shaukat, Z., Aiman, S., Zhou, T. & Li, C. A systems biology-driven approach to construct a comprehensive protein interaction network of influenza A virus with its host. BMC Infectious Diseases 20, 480 (2020). Aguiar, R. S. et al. Molecular alterations in the extracellular matrix in the brains of newborns with congenital Zika syndrome. Science Signaling 13, eaay6736 (2020). Soares-Schanoski, A. et al. Systems analysis of subjects acutely infected with the Chikungunya virus. PLoS Pathog 15, e1007880 (2019). Creighton, R. et al. Network vaccinology. Seminars in Immunology 50, 101420 (2020). Kwissa, M. et al. Dengue Virus Infection Induces Expansion of a CD14 + CD16 + Monocyte Population that Stimulates Plasmablast Differentiation. Cell Host & Microbe 16, 115–127 (2014). Sekaly, R. & Pulendran, B. Systems biology in understanding HIV pathogenesis and guiding vaccine development. Curr Opin HIV AIDS 7, 1–3 (2012). Rao, S., Ghosh, D., Asturias, E. J. & Weinberg, A. What can we learn about influenza infection and vaccination from transcriptomics? Hum Vaccin Immunother 15, 2615–2623 (2019). Winter, C., Camarão, A. A. R., Steffen, I. & Jung, K. Network meta-analysis of transcriptome expression changes in different manifestations of dengue virus infection. BMC Genomics 23, 165 (2022). Barrett, T. et al. NCBI GEO: archive for functional genomics data sets—update. Nucleic Acids Research 41, D991–D995 (2013). Russo, P. S. T. et al. CEMiTool: a Bioconductor package for performing comprehensive modular co-expression analyses. BMC Bioinformatics 19, 56 (2018). Davis, S. & Meltzer, P. S. GEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor. Bioinformatics 23, 1846–1847 (2007). Kauffmann, A., Gentleman, R. & Huber, W. arrayQualityMetrics—a bioconductor package for quality assessment of microarray data. Bioinformatics 25, 415–416 (2009). Howe, K. L. et al. Ensembl 2021. Nucleic Acids Research 49, D884–D891 (2021). Gonçalves, A. N. A. et al. Assessing the Impact of Sample Heterogeneity on Transcriptome Analysis of Human Diseases Using MDP Webtool. Front Genet 10, 971 (2019). Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, i884–i890 (2018). Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013). Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010). Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43, e47 (2015). MetaVolcanoR. Bioconductor http://bioconductor.org/packages/MetaVolcanoR/ . Yu, G., Wang, L.-G., Han, Y. & He, Q.-Y. clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters. OMICS 16, 284–287 (2012). Li, S. et al. Molecular signatures of antibody responses derived from a systems biology study of five human vaccines. Nat Immunol 15, 195–204 (2014). GeneOverlap. Bioconductor http://bioconductor.org/packages/GeneOverlap/ . de Lima, D. S. et al. Long noncoding RNAs are involved in multiple immunological pathways in response to vaccination. Proc Natl Acad Sci U S A 116, 17121–17126 (2019). McNab, F., Mayer-Barber, K., Sher, A., Wack, A. & O’Garra, A. Type I interferons in infectious disease. Nat Rev Immunol 15, 87–103 (2015). Additional Declarations No competing interests reported. Supplementary Files TableS1metaDEGs.xlsx TableS2consensusNetwork.xlsx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 23 Jun, 2025 Reviews received at journal 18 Jun, 2025 Reviews received at journal 12 Jun, 2025 Reviews received at journal 12 Jun, 2025 Reviewers agreed at journal 28 May, 2025 Reviewers agreed at journal 28 May, 2025 Reviewers agreed at journal 25 May, 2025 Reviewers agreed at journal 24 May, 2025 Reviewers agreed at journal 22 May, 2025 Reviewers invited by journal 22 May, 2025 Editor assigned by journal 12 May, 2025 Submission checks completed at journal 08 May, 2025 First submitted to journal 10 Oct, 2024 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-5240411","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":530208143,"identity":"332dda57-a933-4f20-af69-cd697ca36c31","order_by":0,"name":"Vanessa Escolano Maso","email":"","orcid":"","institution":"University of São Paulo","correspondingAuthor":false,"prefix":"","firstName":"Vanessa","middleName":"Escolano","lastName":"Maso","suffix":""},{"id":530208144,"identity":"08cb8ef2-d796-4e50-ac32-8ac09d13435b","order_by":1,"name":"Ana Paula Barbosa","email":"","orcid":"","institution":"Cincinnati Children's Hospital Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Ana","middleName":"Paula","lastName":"Barbosa","suffix":""},{"id":530208145,"identity":"ce7eb568-9499-4217-b74b-bb9a9478de85","order_by":2,"name":"Viviane Schuch","email":"","orcid":"","institution":"Morehouse School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Viviane","middleName":"","lastName":"Schuch","suffix":""},{"id":530208146,"identity":"d04be4f5-195f-4741-8409-1ec17504df77","order_by":3,"name":"Thomas Hagan","email":"","orcid":"","institution":"Cincinnati Children's Hospital Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Hagan","suffix":""},{"id":530208148,"identity":"7b5b8a38-a822-40de-a441-268439b3400f","order_by":4,"name":"Helder Nakaya","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAqUlEQVRIiWNgGAWjYBACAwYGZhAtByKYSdJiTLqWxAaitZhLNz82+Nlml77h/NkDzIV7iNBiOeeYcWJvW3Luhht5CcwznhHjsBsJxgd425iBWngMmHkOEKUl/fPBv2316QbnzxCtJcc4mbftcILBgRzitRQby5w7bjgT6JfDM4h02GbJN2XV8nznzx58XECMFjBgZAORPAxEawCCPxAto2AUjIJRMAqwAgBfFjlOpgUY4wAAAABJRU5ErkJggg==","orcid":"","institution":"University of São Paulo","correspondingAuthor":true,"prefix":"","firstName":"Helder","middleName":"","lastName":"Nakaya","suffix":""}],"badges":[],"createdAt":"2024-10-10 14:23:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5240411/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5240411/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94887572,"identity":"c6a03454-4a9a-4c04-a8bf-61e95cbf06f8","added_by":"auto","created_at":"2025-10-31 18:56:30","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6836,"visible":true,"origin":"","legend":"","description":"","filename":"45e1af517e2648ec85ff9415a75717e5.json","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/98db443b54c2b02e086869cd.json"},{"id":94986918,"identity":"6add89d5-04b8-4002-9d6a-2c25f9a2adcb","added_by":"auto","created_at":"2025-11-03 07:00:58","extension":"html","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":85301,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/0a088af46ac67aec93a92e15.html"},{"id":94887571,"identity":"3e123128-8396-4df1-a023-30eb690496a6","added_by":"auto","created_at":"2025-10-31 18:56:30","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":196992,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMeta-analysis of transcriptomic datasets of viral infections.\u003c/strong\u003e (A) Overview of the number of studies and samples analyzed for Yellow Fever Virus (YFV), Chikungunya Virus (CHIKV), Dengue Virus (DENV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU). (B) Workflow of the analytical process from raw data to functional analysis, including normalization, outlier detection using the Molecular Degree of Perturbation (MDP) method, and the identification and analysis of differentially expressed genes (DEGs) and co-expression patterns. (C) Molecular Degree of Perturbation (MDP) scores for samples in the Dengue study (GSE28988), highlighting outliers by red arrows. (D) Venn diagram demonstrating that removal of outliers in the Dengue study improves DEG analysis by revealing biologically relevant pathways and genes, such as those involved in the T cell receptor signaling pathway (WP69).\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/6e319f313eb42190c217c6ea.jpg"},{"id":94887573,"identity":"1d6620c3-6be6-40b3-9bc8-868f4dffbb43","added_by":"auto","created_at":"2025-10-31 18:56:30","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":454942,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDifferentially expressed genes (DEGs) and meta-analysis across viral infections. (A) \u003c/strong\u003eBox plots showing the number of up-regulated and down-regulated DEGs in each study for Dengue Virus (DENV), Chikungunya Virus (CHIKV), Yellow Fever Virus (YFV), Influenza Virus (FLU), and Human Immunodeficiency Virus (HIV). Circles represent studies using microarray technology, and triangles represent studies using RNA-seq technology.\u003cstrong\u003e (B) \u003c/strong\u003eBar plots displaying the number of meta-DEGs identified after meta-analysis using Fisher's method (meta-FDR \u0026lt; 0.05, median log2 fold-change \u0026gt; 1), with up-regulated genes in red and down-regulated genes in blue for each virus.\u003cstrong\u003e(C) \u003c/strong\u003ePathway enrichment analysis (over-representation analysis using blood transcription modules) of the meta-DEGs, represented by dot plots showing enriched pathways for each virus. Dot size and color gradient correspond to the gene ratio and -log10 (adjusted p-value), respectively. Colors indicate different cell types or functions.\u003cstrong\u003e (D) \u003c/strong\u003eNetwork of up-regulated meta-DEGs, with nodes and edges colored by Louvain modules. Highlighted are genes shared by all viruses and genes shared by arboviruses (DENV, CHIKV, YFV).\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/8f2c3f881d12c183a75d0a31.jpg"},{"id":94887574,"identity":"d204e10e-9980-468f-9071-be4ca8074f3e","added_by":"auto","created_at":"2025-10-31 18:56:30","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":272546,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConsensus network analysis of co-expression modules across different viral infections. (A) \u003c/strong\u003eBar plot showing the number of genes included in the consensus co-expression network for each virus: Dengue Virus (DENV), Chikungunya Virus (CHIKV), Yellow Fever Virus (YFV), Influenza Virus (FLU), and Human Immunodeficiency Virus (HIV).\u003cstrong\u003e (B) \u003c/strong\u003eBar plot displaying the number of community modules identified in the consensus network for each virus.\u003cstrong\u003e (C) \u003c/strong\u003eVisualization of the Dengue Virus (DENV) consensus co-expression network, highlighting 18 community modules (C0 to C17) in different colors. \u003cstrong\u003e(D) \u003c/strong\u003eIdentification of biological process or cell-type related genes within the community modules. The color of the gene symbols corresponds to the color of the community module in which they are found, indicating processes such as neutrophil activity, plasmablast presence, platelet function, T cell response, inflammation and antiviral response, and complement system involvement.\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/a9661083390aea28e28d59f7.jpg"},{"id":94887575,"identity":"0b9b1813-d770-46ab-83fa-9f78b0780925","added_by":"auto","created_at":"2025-10-31 18:56:30","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":324131,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverlap of co-expression community modules across different viral infections and associated biological processes. (A) \u003c/strong\u003eHeatmap showing the overlap of genes between community modules from different viruses, with color intensity proportional to the Jaccard index. Only squares with significant overlaps are displayed (Adjusted P \u0026lt; 0.05, Fisher's exact test). The virus and community module numbers are indicated. Notable overlaps are highlighted in black, brown, blue, and green.\u003cstrong\u003e (B) \u003c/strong\u003eList of genes from the highlighted black squares in panel A associated with the antiviral response.\u003cstrong\u003e(C) \u003c/strong\u003eList of genes from the highlighted green squares in panel A associated with plasma B cells and plasmablasts. \u003cstrong\u003e(D) \u003c/strong\u003eList of genes from the highlighted brown and blue squares in panel A associated with hemoglobin, heme metabolism, and erythropoiesis, and with platelets, respectively.\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/52793b3cc3f33842b473a3ee.jpg"},{"id":95000650,"identity":"b51664a0-4445-41ce-a792-f03365e61663","added_by":"auto","created_at":"2025-11-03 08:59:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1923470,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/52e603e9-03d1-4a18-8456-ebb5e5739909.pdf"},{"id":94987031,"identity":"faaf6091-ed8d-4b04-a91d-88c407f77302","added_by":"auto","created_at":"2025-11-03 07:01:07","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":239416,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1metaDEGs.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/10044a7f268a3acaf41a37f0.xlsx"},{"id":94887578,"identity":"e6a45026-49e9-4b3f-8312-ae8f83ebeec5","added_by":"auto","created_at":"2025-10-31 18:56:30","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":339913,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2consensusNetwork.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5240411/v1/b6eaf2300a6146d7e75cc048.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Network Medicine Approach to Viral Infections: Unraveling Universal and Virus-Specific Immune Pathways","fulltext":[{"header":"Introduction","content":"\u003cp\u003eViral infections remain a significant global health challenge, contributing to widespread morbidity and mortality across the world\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Pathogens such as Dengue Virus (DENV), Chikungunya Virus (CHIKV), Yellow Fever Virus (YFV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU) infect millions of people annually, causing a wide range of clinical outcomes and posing an ongoing threat to public health systems\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. These viruses have distinct epidemiological characteristics and transmission routes, with symptoms that can vary from mild to life-threatening. For instance, DENV can lead to severe hemorrhagic manifestations\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, while CHIKV often causes chronic arthralgia\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e, and YFV is associated with potentially fatal liver failure\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Understanding the host immune response to these infections is essential for the development of more effective diagnostics, treatments, and vaccines. However, the molecular mechanisms that drive these immune responses, particularly the differentiation between shared and virus-specific pathways, are not fully understood, making this a critical area of research for improving the management of viral diseases.\u003c/p\u003e \u003cp\u003eSystems biology and network medicine approaches have transformed our understanding of host-pathogen interactions by emphasizing the interconnectedness of molecular processes during viral infections\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Rather than focusing on individual genes, these approaches map the complex interactions between genes, proteins, and cellular pathways, offering a more holistic view of how the immune system responds to viral threats. Transcriptomic studies, which capture genome-wide gene expression profiles, are essential tools for constructing these interaction networks\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. By identifying differentially expressed genes (DEGs), researchers can pinpoint key nodes and pathways that play pivotal roles in immune responses, disease progression, and recovery. Analysis of whole blood and peripheral blood mononuclear cells (PBMCs) during viral infections has revealed crucial genes and pathways involved in antiviral defense, inflammation, and immune modulation\u003csup\u003e\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. However, while transcriptomic data is valuable for generating these molecular networks, individual studies often lack the statistical power to provide a comprehensive understanding of the system\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Thus, integrating data across multiple studies and viral infections is crucial to uncover both shared immune mechanisms and virus-specific responses, ultimately revealing conserved molecular patterns that can inform broad-spectrum antiviral therapies and identify unique pathways for virus-specific treatments.\u003c/p\u003e \u003cp\u003eDespite the vast availability of transcriptomic datasets, comparing gene expression patterns across different viruses using a systems biology approach remains challenging. Variations in study design, sample sizes, and the use of different technologies, such as microarray and RNA-seq, can introduce biases that complicate the construction of accurate molecular networks. Moreover, biological heterogeneity among infected individuals further complicates the interpretation of network structures, as variations in age, genetic background, or comorbidities can obscure consistent molecular signatures. To our knowledge, no study has systematically employed a transcriptomic meta-analysis across multiple viral infections to identify consistent molecular signatures and network structures. A network medicine approach, integrating data from different viruses, holds the potential to reveal both common and virus-specific molecular networks, offering deeper insights into immune responses and uncovering potential therapeutic targets.\u003c/p\u003e \u003cp\u003eIn this study, we applied a systems biology approach to perform a meta-analysis of transcriptomic datasets from whole blood and PBMC samples of individuals acutely infected with DENV, CHIKV, YFV, HIV, and FLU. Our primary objective was to construct robust molecular networks that capture the immune responses to these viral infections by integrating data from multiple studies and platforms. By leveraging these networks, we sought to uncover both shared and virus-specific molecular mechanisms and pathways. Through pathway enrichment analyses, we further elucidated the biological processes and immune responses activated by each virus. Our findings contribute to a deeper understanding of the universal and virus-specific immune pathways involved in viral infections, offering potential avenues for therapeutic interventions, and enhancing our knowledge of host-pathogen interactions in the context of multiple viral threats.\u003c/p\u003e"},{"header":"Material and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Selection and Data Collection\u003c/h2\u003e \u003cp\u003eThe studies were identified through systematic searches in the Gene Expression Omnibus (GEO) database\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. The following six inclusion criteria were applied: 1) samples from humans in the acute phase of infection by one of the five viruses\u0026mdash;DENV, YFV, CHIKV, FLU, or HIV\u0026mdash;as well as samples from healthy individuals and/or those in the convalescent phase of viral infection; 2) sample donors must not have received any medicinal treatments; 3) samples must be obtained from whole blood or peripheral blood mononuclear cells (PBMCs); 4) transcriptomic studies must have been conducted using microarray or bulk RNA-seq technologies; 5) sequencing platforms must belong to either Illumina or Affymetrix; 6) a minimum of 20 samples must be available after excluding outliers. The minimum sample size was determined based on the number required to construct robust co-expression networks, meaning networks whose topology remains stable as additional samples are include\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. Studies meeting all six inclusion criteria were incorporated into our meta-analysis.\u003c/p\u003e \u003cp\u003eWe selected transcriptomic datasets from both microarray and RNA-seq platforms for five viral infections or vaccination with attenuated virus: DENV, YFV, CHIKV, Influenza Virus (FLU), and HIV. For DENV, six microarray studies (GSE13052, GSE25001, GSE28405, GSE28988, GSE29891, GSE51808) and one RNA-seq study (GSE157240) were included, comprising a total of 532 infected and 212 non-infected control samples. The YFV dataset comprises of PBMC from 14 convalescent and 41 acutely infected individuals (GSE243442) using RNA-seq.\u0026nbsp;For CHIKV, two RNA-seq studies (GSE99992, PRJNA507472) contributed a total of 81 infected and 60 control samples. FLU datasets included two microarray studies (GSE29366, GSE34205) and one RNA-seq study (GSE157240), totaling 154 samples (112 infected, 42 control). For HIV, two microarray datasets (GSE29429 using the GPL10558 platform and GSE29429 using the GPL6947 platform) were included, providing 56 infected and 53 control samples.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eData processing and Normalization\u003c/h3\u003e\n\u003cp\u003eRaw microarray gene expression data were downloaded from GEO (\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) using the GEOquery package\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Data were imported into RStudio using the ReadAffy function from the affy package for Affymetrix platforms and the read.ilmn function from the limma package for Illumina platforms. Quality control was conducted using the arrayQualityMetrics function from the arrayQualityMetrics package\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Background correction, quantile normalization, and log2 transformation were performed using the neqc function from the limma package for Illumina data and the justRMA function from affy for Affymetrix data. Quality control of normalized data included identifying and removing outlier samples using the arrayQualityMetrics package. Normalized data were further scrutinized for outliers, defined as samples flagged by at least two of the following tests: i) mean absolute difference of M-values (log-ratios) between arrays, ii) the Kolmogorov-Smirnov statistic (Ka) comparing each array\u0026rsquo;s signal distribution to pooled data, and iii) Hoeffding\u0026rsquo;s statistic (Da) on the joint distribution of A (average) and M values. Probes were annotated at the gene level according to the human genome version GRCh38.p13, Ensembl 104\u003csup\u003e17\u003c/sup\u003e, using the reannotator_microarray_probes pipeline, which aligns probe nucleotide sequences to the reference genome. Summarization retained the probe with the highest mean expression for each gene. Phenotypic outliers were identified using the Molecular Degree of Perturbation (MDP) package\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e, which evaluates the expression of the top 25% most perturbed genes. Identified outliers were excluded from subsequent analyses.\u003c/p\u003e \u003cp\u003eRaw RNA-seq data were obtained from the Sequence Read Archive (SRA) using prefetch from the SRA Toolkit. Metadata for each study were retrieved using the SRA Run Selector. The SRA files were converted to fastq format with fasterq-dump from the SRA Toolkit. Quality control, adapter removal, and filtering of low-quality reads were performed using the fastp package\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. For paired-end reads, the parameters --detect_adapter_for_pe, --cut_front, --cut_right, --qualified_quality_phred 20, --correction, and --overrepresentation_analysis were used. The same parameters were applied for single-end reads, except --detect_adapter_for_pe and --correction, which are specific to paired-end data. Reads were aligned to the human genome (GRCh38.p13, Ensembl release 104) using the STAR package\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Genomic indices were constructed using the primary assembly FASTA file and the corresponding GTF annotation file with the --sjdbOverhang 100 parameter. Aligned reads were stored in BAM files and sorted by genomic coordinates. Read summarization was conducted using featureCounts. For paired-end data, only read pairs aligned to the genome were counted. Chimeric fragments, multi-mapping, and multi-overlapping reads were excluded. Summarization was performed at the gene level using Ensembl gene identifiers (ENSG), considering strandedness according to the methodology reported in each study. Genes with zero counts or low expression were removed from the count table. Low expression was defined as fewer than one count per million (CPM) in at least x samples, where x is the sample size of the smallest phenotypic group. Library size was calculated by summing the counts of retained genes. Library size was normalized using the Trimmed Mean of M-values (TMM) method implemented in the calcNormFactors function of the edgeR package\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. The count table was normalized using the CPM method implemented by the cpm function from edgeR. Biological outliers in the RNA-seq data were identified using the MDP test, which assesses the expression of the 25% most perturbed genes. Identified outliers were excluded from subsequent analyses.\u003c/p\u003e\n\u003ch3\u003eDifferential Gene Expression Meta-Analysis\u003c/h3\u003e\n\u003cp\u003eFor individual microarray studies, differentially expressed genes (DEGs) were identified using the limma package in R\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, which employs a moderated t-test. For bulk RNA-seq studies, differential expression analysis was performed using the edgeR package. DEG analysis was conducted using the quasi-likelihood F-test, recommended for bulk RNA-seq data with biological replicates. DEGs were identified in each study by comparing samples from the acute phase of infection with control group samples, which included healthy and/or convalescent individuals.\u003c/p\u003e \u003cp\u003eMeta-analysis was performed using Fisher's method for combining p-values, implemented in the MetaVolcanoR package\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. A meta-analysis was conducted for each acute viral infection studied: DENV, CHIKV, FLU, and HIV. However, meta-analysis for YFV was not possible due to the availability of only one human blood transcriptome study for this infection. The p-values obtained from the meta-analysis were adjusted for multiple testing using the Benjamini-Hochberg False Discovery Rate (FDR-BH) method. Log2 fold change (log2FC) was calculated as the median log2FC for each gene across studies of the same infection. Genes were considered consistently differentially expressed for infection if they had an FDR-BH\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and |log2FC| \u0026gt; 1.\u003c/p\u003e\n\u003ch3\u003eDisease-gene networks\u003c/h3\u003e\n\u003cp\u003eTwo disease-gene networks were constructed: one for consistently up-regulated genes and another for consistently down-regulated genes. The networks were built using Gephi software. Community detection in each network was performed using the Louvain algorithm. For network visualization, nodes were colored according to their community, with each community represented by a distinct color. The ForceAtlas algorithm was applied to spatialize the networks.\u003c/p\u003e \u003cp\u003eOver-representation analysis (ORA) was conducted separately for the lists of consistently up-regulated and down-regulated genes from each infection analyzed in this study: DENV, CHIKV, FLU, and HIV. For YFV, the list of DEGs was used instead, due to the availability of only one dataset. ORA was performed using the clusterProfiler package\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, which applies a right-tailed Fisher\u0026rsquo;s exact test. Blood Transcription Modules (BTMs) were used as the reference gene set, as they represent groups of genes previously identified as expressed in blood samples\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Gene sets were considered enriched if they had an FDR-BH\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e\n\u003ch3\u003eCoexpression consensus network construction\u003c/h3\u003e\n\u003cp\u003eGene coexpression networks and their modules were constructed using the CEMiTool R package\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e with the normalized gene expression matrix after removing outlier samples. Pearson positive correlations were used to identify modules. In bulk RNA-seq studies, the dependence between the mean and variance of gene expression was assessed. When such dependence was detected, the normalized gene expression matrix was transformed using the Variance Stabilizing Transformation (VST) method.\u003c/p\u003e \u003cp\u003eA consensus coexpression network was constructed for each acute viral infection by creating a list of edges from the coexpression network of each study, then comparing these lists across all studies for the same infection, retaining only the edges present in more than half of the studies. Networks were visualized using Gephi software with the ForceAtlas2 algorithm. Community detection was performed using the Louvain algorithm. Centrality measures for the nodes, including degree, betweenness centrality, and harmonic centrality, were calculated. The biological functions associated with each community in the consensus coexpression networks were identified through ORA functional enrichment analysis, implemented with clusterProfiler\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, using BTMs as reference gene sets. Enriched BTMs were identified based on an FDR-BH\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003cp\u003eThe communities in the consensus coexpression networks for DENV, CHIKV, FLU, and HIV, as well as the coexpression modules for YFV, were compared to identify similar or unique communities among the viral infections. This comparison was performed using the GeneOverlap package\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, which conducts a right-tailed Fisher\u0026rsquo;s exact test to assess the overlap between two gene sets, with each community treated as a gene set. Additionally, the Jaccard index was calculated using GeneOverlap to measure the similarity between community pairs.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eMeta-Analysis of Transcriptomic Datasets Across Different Viral Infections\u003c/h2\u003e \u003cp\u003eWe analyzed transcriptomic datasets of blood and peripheral blood mononuclear cells (PBMC) from individuals acutely infected with five different viruses: Yellow Fever Virus (YFV), Chikungunya Virus (CHIKV), Dengue Virus (DENV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU). Our analysis aimed to identify consistent and robust signatures of viral infection and to reveal potential molecular mechanisms shared or specific to each infection. The number of studies and samples included in our analysis varied by virus. We included one study for YFV (41 infected and 14 control samples), two studies for CHIKV (80 infected and 60 control samples), seven studies for DENV (316 infected and 213 control samples), two studies for HIV (56 infected and 53 control samples), and three studies for FLU (112 infected and 42 control samples) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Our analytical workflow consisted of several key steps (see methods). We started with the collection of raw transcriptomic data, followed by quality control and normalization of gene expression levels. We then employed the Molecular Degree of Perturbation (MDP) method\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e to detect outliers. Subsequent steps included the identification of differentially expressed genes (DEGs) and co-expression patterns, followed by functional analysis. This workflow ensured the robustness and reliability of our meta-analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003eTo address sample heterogeneity, we applied the MDP method to all transcriptomic datasets. MDP quantifies the perturbation score of samples using nonperturbed subjects (i.e., healthy controls) as a reference group. This score allows the detection of outlying samples and subgroups of diseased patients, which is crucial because individual heterogeneity can significantly impact gene expression analyses. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC and D illustrate a single example of this improvement using the Dengue virus study (GSE28988). The MDP scores for control samples (light blue) and infected samples (black) revealed several potential outliers (shown in red). Identifying and removing these outliers was crucial, as they can significantly impact the accuracy of DEG analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC). For instance, the removal of outliers in GSE28988 dataset unveiled the T cell receptor signaling pathway (WP69), including genes such as MAP4K1, ITK, CD247, CD3E, ICOS, and SKAP1, which were not detected when outliers were included (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). This underscores the importance of addressing sample heterogeneity to uncover biologically relevant pathways and genes.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNext, we conducted a comprehensive meta-analysis of transcriptomic datasets from blood and PBMC samples of individuals acutely infected with these five different viruses. This analysis aimed to identify consistent and robust signatures of viral infection and uncover potential molecular mechanisms that are either shared across or specific to different infections. The number of up-regulated and down-regulated DEGs identified in each study for the five viruses varied across different studies (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Notably, DENV and CHIKV had a higher number of up-regulated genes compared to the other viruses (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003eWe then performed a meta-analysis of DEGs using Fisher's method (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). This analysis aggregated the DEGs from individual studies to identify meta-DEGs, representing consistently differentially expressed genes across multiple studies (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). Again, CHIKV and DENV showed the highest number of meta-DEGs, underscoring the robustness of these signatures across different datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). Pathway enrichment analysis was conducted on the meta-DEGs using over-representation analysis with blood transcription modules (BTM) as gene sets (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). This analysis revealed distinct pathway enrichment profiles for each virus, providing insights into the specific immune responses and biological processes activated during infection (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003eTo check for shared and specific meta-DEGs between the different viruses, we created a network using up-regulated meta-DEGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). This network highlights genes that are shared among all viruses and those specific to arboviruses (DENV, CHIKV, YFV). Genes such as SPATS2L, OASL, and HERC6, which are shared by multiple viruses, may play crucial roles in the common antiviral response, while other genes are unique to specific viral infections, suggesting virus-specific pathways and mechanisms (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). For instance, TNFSF10, PLAC8, and NT5C3A are uniquely up-regulated in arboviruses like DENV, CHIKV, and YFV, indicating virus-specific pathways and mechanisms (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). This distinction between shared and unique genes provides insight into both universal and virus-specific immune responses.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo further explore the molecular mechanisms underlying different viral infections, we performed a consensus network analysis of co-expression modules across the transcriptomic datasets\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. This analysis aimed to identify genes and pathways that are consistently co-expressed across different viruses, thereby highlighting potential shared and virus-specific mechanisms. The number of genes included in the consensus co-expression network for each virus varied significantly, with YFV having the highest number of genes (2,223) and DENV the lowest (397) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). The number of community modules identified within the consensus network also varied, with CHIKV having the highest number of modules (31) and YFV the lowest (8) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe DENV consensus co-expression network highlights 18 community modules (C0 to C17), each represented by a different color (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). These modules represent distinct groups of genes that are co-expressed during DENV infection, indicating coordinated biological processes. We identified various biological processes and cell-type related genes within these community modules. For example, genes related to neutrophil activity (e.g., LTF, DEFA1, CTSG, AZU1, MPO) and plasmablast presence (e.g., CD38, XBP1, TNFRSF17, JCHAIN, MZB1, CXCR3, TOP2A) were found in specific modules. Other modules included genes related to platelet function (e.g., ITGA2B, PF4V1, ITGB3, CLEC5A, CLEC1B, CLEC4F, CLEC12A), T cell response (e.g., CD8A, GZMK, GZMA, IL7R, CCR7, CXCR5), inflammation and antiviral response (e.g., TLR7, TLR4, ISG15, OAS1, OAS2, OAS3, OASL, IFIT1, IFIT2, IFIT3, IFIT5, MX1, MX2, CXCL10, GBP1, GBP5, GBP6, GBP4, GBP3), and complement system involvement (e.g., C4BPA, C1QA, C1QB, C1QC, C3AR1, SERPING1) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo identify shared and virus-specific co-expression patterns, we analyzed the overlap of genes between community modules from different viruses. The overlap analysis revealed significant co-expression patterns across the viral infections, with notable overlaps highlighted. For example, the black highlighted overlaps included genes related to the antiviral response such as ATF3, CCL2, and CXCL10 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). The green highlighted overlaps included genes associated with plasma B cells and plasmablasts, such as CD27, CD38, and CXCR3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Additionally, the brown and blue highlighted overlaps revealed genes involved in hemoglobin, heme metabolism, erythropoiesis (e.g., AHSP, HBA1, HBB, SLC25A39), and platelet function (e.g., ALOX12, ITGA2B, GP6, GP9) respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we identified several consistently up-regulated genes across different viral infections, such as OAS2, MX1, and IFI44L, which, although commonly associated with the interferon response\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e, highlight additional layers of complexity in antiviral defense mechanisms. These genes are part of broader pathways involved not only in direct antiviral activity but also in modulating immune responses, inflammation, and cellular stress. For example, OAS2 has roles beyond RNA degradation, influencing immune signaling cascades, while MX1 has been shown to interact with other host proteins to limit viral replication. These findings underscore the importance of these genes in coordinating a more nuanced response to viral infections and highlight their potential as biomarkers for disease severity or progression. Furthermore, the expression of genes like IFI44L suggests possible roles in modulating the immune environment, contributing to the differentiation between acute and chronic infection stages across multiple viral threats. This suggests that while interferon response is a common theme, these genes may also be involved in more specialized functions that tailor immune responses to specific viral challenges.\u003c/p\u003e \u003cp\u003eWhile we found shared molecular signatures, several virus-specific pathways also emerged from our analysis. For instance, genes like TNFSF10 and NT5C3A were up-regulated in arboviruses such as DENV, CHIKV, and YFV, but not in HIV or FLU. This suggests that certain immune pathways are more relevant to particular viral families, possibly reflecting unique interactions between the virus and host immune system. Pathway enrichment analysis further supported this, with pathways related to immune regulation, such as cytokine signaling and T cell receptor signaling, being more pronounced in specific viral infections. These findings open opportunities for virus-specific therapeutic interventions, as targeting these unique pathways could provide more tailored treatments for diseases like DENV and CHIKV.\u003c/p\u003e \u003cp\u003eDespite these robust findings, there were several challenges in our analysis. Variations in study design, differences in sample sizes, and the use of both microarray and RNA-seq technologies introduced potential biases. Furthermore, biological heterogeneity within patient populations, including factors such as age, genetic background, and co-morbidities, complicated the identification of consistent molecular signatures. Although the MDP method allowed us to address some of this heterogeneity by identifying and removing outliers, these complexities underscore the need for more standardized datasets and methodologies in transcriptomic meta-analyses. Additionally, while our co-expression network analysis uncovered significant shared and virus-specific pathways, the extent to which these findings can be generalized across other viral infections requires further investigation.\u003c/p\u003e \u003cp\u003eOur study provides valuable insights into both shared and virus-specific immune responses to viral infections, contributing to the broader understanding of host-pathogen interactions. The identification of conserved molecular networks across different viruses offers potential targets for broad-spectrum antiviral therapies, while the virus-specific pathways suggest avenues for more personalized treatments. Moving forward, expanding the analysis to include additional viral infections and larger datasets will further strengthen the understanding of these immune mechanisms. Future research should also focus on integrating multi-omic data, such as proteomics and metabolomics, with transcriptomic data to provide a more comprehensive view of host responses to viral infections. Additionally, the development of standardized protocols for transcriptomic studies could help mitigate the technical biases observed in our analysis, ultimately enhancing the reproducibility and robustness of meta-analyses in systems biology.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eDisclosures\u003c/h2\u003e \u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis work was supported by the S\u0026atilde;o Paulo State Research Foundation (FAPESP) (grant numbers: 2018/14933-2 to H.I.N; 2023/05005-2 and 2019/16418-0 to V.E.M.).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization: V.E.M., H.I.N.; Data curation: All authors; Formal analysis: V.E.M., A.P.B.N.O., V.S., D.S.L.; Funding acquisition: H.I.N.; Methodology: All authors; Writing \u0026ndash; original draft: V.E.M., H.I.N.; Writing \u0026ndash; review and editing: All authors; Supervision: T.H., H.I.N.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThe authors would like to thank all the members of Computational Systems Biology Laboratory (CSBL) for their input.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eNaghavi, M. \u003cem\u003eet al.\u003c/em\u003e Global burden associated with 85 pathogens in 2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet Infectious Diseases 24, 868\u0026ndash;895 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKularatne, S. A. \u0026amp; Dalugama, C. Dengue infection: Global importance, immunopathology and management. Clin Med (Lond) 22, 9\u0026ndash;13 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchilte, C. \u003cem\u003eet al.\u003c/em\u003e Chikungunya Virus-associated Long-term Arthralgia: A 36-month Prospective Longitudinal Study. PLOS Neglected Tropical Diseases 7, e2137 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKallas, E. G. \u003cem\u003eet al.\u003c/em\u003e Predictors of mortality in patients with yellow fever: an observational cohort study. The Lancet Infectious Diseases 19, 750\u0026ndash;758 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarooq, Q. ul A., Shaukat, Z., Aiman, S., Zhou, T. \u0026amp; Li, C. A systems biology-driven approach to construct a comprehensive protein interaction network of influenza A virus with its host. BMC Infectious Diseases 20, 480 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAguiar, R. S. \u003cem\u003eet al.\u003c/em\u003e Molecular alterations in the extracellular matrix in the brains of newborns with congenital Zika syndrome. Science Signaling 13, eaay6736 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoares-Schanoski, A. \u003cem\u003eet al.\u003c/em\u003e Systems analysis of subjects acutely infected with the Chikungunya virus. PLoS Pathog 15, e1007880 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCreighton, R. \u003cem\u003eet al.\u003c/em\u003e Network vaccinology. Seminars in Immunology 50, 101420 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwissa, M. \u003cem\u003eet al.\u003c/em\u003e Dengue Virus Infection Induces Expansion of a CD14\u0026thinsp;+\u0026thinsp;CD16\u0026thinsp;+\u0026thinsp;Monocyte Population that Stimulates Plasmablast Differentiation. Cell Host \u0026amp; Microbe 16, 115\u0026ndash;127 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSekaly, R. \u0026amp; Pulendran, B. Systems biology in understanding HIV pathogenesis and guiding vaccine development. Curr Opin HIV AIDS 7, 1\u0026ndash;3 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRao, S., Ghosh, D., Asturias, E. J. \u0026amp; Weinberg, A. What can we learn about influenza infection and vaccination from transcriptomics? Hum Vaccin Immunother 15, 2615\u0026ndash;2623 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWinter, C., Camar\u0026atilde;o, A. A. R., Steffen, I. \u0026amp; Jung, K. Network meta-analysis of transcriptome expression changes in different manifestations of dengue virus infection. BMC Genomics 23, 165 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarrett, T. \u003cem\u003eet al.\u003c/em\u003e NCBI GEO: archive for functional genomics data sets\u0026mdash;update. Nucleic Acids Research 41, D991\u0026ndash;D995 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRusso, P. S. T. \u003cem\u003eet al.\u003c/em\u003e CEMiTool: a Bioconductor package for performing comprehensive modular co-expression analyses. BMC Bioinformatics 19, 56 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavis, S. \u0026amp; Meltzer, P. S. GEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor. Bioinformatics 23, 1846\u0026ndash;1847 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKauffmann, A., Gentleman, R. \u0026amp; Huber, W. arrayQualityMetrics\u0026mdash;a bioconductor package for quality assessment of microarray data. Bioinformatics 25, 415\u0026ndash;416 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHowe, K. L. \u003cem\u003eet al.\u003c/em\u003e Ensembl 2021. Nucleic Acids Research 49, D884\u0026ndash;D891 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGon\u0026ccedil;alves, A. N. A. \u003cem\u003eet al.\u003c/em\u003e Assessing the Impact of Sample Heterogeneity on Transcriptome Analysis of Human Diseases Using MDP Webtool. Front Genet 10, 971 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, S., Zhou, Y., Chen, Y. \u0026amp; Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, i884\u0026ndash;i890 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDobin, A. \u003cem\u003eet al.\u003c/em\u003e STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15\u0026ndash;21 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobinson, M. D., McCarthy, D. J. \u0026amp; Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139\u0026ndash;140 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRitchie, M. E. \u003cem\u003eet al.\u003c/em\u003e limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43, e47 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMetaVolcanoR. \u003cem\u003eBioconductor\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://bioconductor.org/packages/MetaVolcanoR/\u003c/span\u003e\u003cspan address=\"http://bioconductor.org/packages/MetaVolcanoR/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, G., Wang, L.-G., Han, Y. \u0026amp; He, Q.-Y. clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters. \u003cem\u003eOMICS\u003c/em\u003e 16, 284\u0026ndash;287 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, S. \u003cem\u003eet al.\u003c/em\u003e Molecular signatures of antibody responses derived from a systems biology study of five human vaccines. Nat Immunol 15, 195\u0026ndash;204 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGeneOverlap. \u003cem\u003eBioconductor\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://bioconductor.org/packages/GeneOverlap/\u003c/span\u003e\u003cspan address=\"http://bioconductor.org/packages/GeneOverlap/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Lima, D. S. \u003cem\u003eet al.\u003c/em\u003e Long noncoding RNAs are involved in multiple immunological pathways in response to vaccination. Proc Natl Acad Sci U S A 116, 17121\u0026ndash;17126 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcNab, F., Mayer-Barber, K., Sher, A., Wack, A. \u0026amp; O\u0026rsquo;Garra, A. Type I interferons in infectious disease. Nat Rev Immunol 15, 87\u0026ndash;103 (2015).\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":"npj-viruses","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [npj Viruses](https://www.nature.com/npjviruses)","snPcode":"44298","submissionUrl":"https://submission.springernature.com/new-submission/44298/3","title":"npj Viruses","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Viral infections, Immune response, Co-expression networks, Transcriptomics, Meta-analysis","lastPublishedDoi":"10.21203/rs.3.rs-5240411/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5240411/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eViral infections continue to pose significant global health challenges, with pathogens like Dengue Virus (DENV), Chikungunya Virus (CHIKV), Yellow Fever Virus (YFV), Human Immunodeficiency Virus (HIV), and Influenza Virus (FLU) contributing to widespread morbidity and mortality. A comprehensive understanding of the host immune response to these infections is crucial for developing more effective diagnostics, treatments, and vaccines. In this study, we conducted a meta-analysis of transcriptomic datasets from whole blood and peripheral blood mononuclear cell (PBMC) samples of individuals infected with DENV, CHIKV, YFV, HIV, and FLU. Utilizing a systems biology approach, we identified both shared and virus-specific molecular mechanisms underlying immune responses to these infections. Several key genes, including OAS2, MX1, and IFI44L, were consistently up-regulated across multiple viral infections, highlighting their roles in antiviral defense, immune modulation, and cellular stress responses. Additionally, virus-specific genes such as TNFSF10 and NT5C3A were found to be uniquely up-regulated in arboviruses, underscoring the distinct immune pathways activated by different viral families. Our co-expression network analysis revealed coordinated gene expression patterns across different viral infections, offering insights into conserved molecular networks that could inform the development of broad-spectrum antiviral therapies. Despite challenges related to biological heterogeneity and variations in study design, this meta-analysis provides a foundation for future research aimed at unraveling the complex immune mechanisms that govern host-pathogen interactions.\u003c/p\u003e","manuscriptTitle":"Network Medicine Approach to Viral Infections: Unraveling Universal and Virus-Specific Immune Pathways","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-31 18:56:25","doi":"10.21203/rs.3.rs-5240411/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-06-23T12:13:41+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-18T14:23:54+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-12T21:14:29+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-12T19:02:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"320173619804575920835808910672228325990","date":"2025-05-28T14:08:40+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"183210044915822116104033043334299209087","date":"2025-05-28T12:45:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"275866372798716979946500733821244554202","date":"2025-05-25T12:24:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"189010445538400019711801146797909463859","date":"2025-05-24T16:00:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"328507834427658936203700184882844488989","date":"2025-05-22T15:48:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-05-22T15:38:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-05-12T08:21:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-05-08T18:07:02+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Viruses","date":"2024-10-10T14:15:00+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-viruses","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [npj Viruses](https://www.nature.com/npjviruses)","snPcode":"44298","submissionUrl":"https://submission.springernature.com/new-submission/44298/3","title":"npj Viruses","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"db21d698-0ae0-4f0f-aab5-07da9018892e","owner":[],"postedDate":"October 31st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":56362921,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":56362922,"name":"Biological sciences/Immunology/Immunogenetics"}],"tags":[],"updatedAt":"2025-10-31T18:56:26+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-31 18:56:25","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5240411","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5240411","identity":"rs-5240411","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}