The research team took advantage of whole blood transcriptomics, a technique that measures the activity of thousands of genes simultaneously in circulating immune cells. Rather than examining individual molecules in isolation, transcriptomics captures the coordinated behaviour of the entire immune system at a single moment, revealing which defensive programmes have been switched on, which have been silenced, and how the composition of the blood cell population itself has shifted in response to infection. This systems-level view is particularly valuable in paediatric respiratory disease, where clinical symptoms overlap considerably between viral and bacterial pathogens and where the biological drivers of severity remain poorly understood.

The study drew on children enrolled in the Drakenstein Child Health Study and the Pneumonia in South Africa Programme, longitudinal research platforms based in South Africa that have followed children in communities where the burden of lower respiratory tract illness is among the highest in the world. The investigators compared whole blood transcriptomes from 127 healthy children with those from 71 children who had mild or asymptomatic SARS-CoV-2 infection, 41 children hospitalised with severe SARS-CoV-2 disease, 47 children hospitalised with respiratory syncytial virus-associated lower respiratory tract illness, and 47 children with pulmonary tuberculosis. This design allowed the researchers to distinguish a general response to respiratory illness from pathogen-specific signatures and, crucially, to identify the molecular features that separate mild from severe COVID-19 in children.

The scale of the transcriptional differences was striking. The team identified more than 5,000 differentially expressed genes across the disease groups, all passing a stringent statistical threshold that controls the false discovery rate at below 5 percent. In children with severe SARS-CoV-2, standout genes included OLFM4, IFI27, CBX7, IGF2BP3 and OTOF. In RSV-associated lower respiratory tract illness, the most distinctive genes were IFI27, OTOF, SIGLEC1, IFI44L and USP18, several of which are well-known markers of interferon-driven antiviral activity. Pulmonary tuberculosis produced its own signature, dominated by MMP8, LTF, IGF2BP3, GPR84, CD177, C1QC and DEFA4, genes associated with neutrophil activation, granule release and the complement cascade, consistent with the intense myeloid inflammation that characterises tuberculous lung disease.

Pathway analysis of the severe COVID-19 group, compared with healthy uninfected children, revealed enrichment of neutrophil degranulation and interferon gamma signalling, alongside overexpression of genes encoding ribosomal proteins and a measurable depletion of general immune response programmes. Neutrophil degranulation refers to the release of toxic granule contents by these frontline white blood cells, a process that helps destroy pathogens but can also damage surrounding tissue when excessive. Interferon gamma, meanwhile, is a cytokine that orchestrates macrophage activation and antiviral defence. The combination of a strong neutrophil-driven inflammatory programme with signs of broader immune dysregulation echoes patterns previously documented in adults with severe COVID-19, and the authors note that severe disease in children exhibits a cellular response similar to that reported in adult patients.

To move beyond lists of individual genes, the researchers applied Weighted Gene Co-expression Network Analysis, a computational method that clusters genes into modules based on correlated patterns of expression across samples. Rather than treating each transcript as an independent variable, WGCNA identifies groups of genes that rise and fall together, which often reflects shared biological regulation. The analysis uncovered 10 such correlated gene modules that were shared across the different lower respiratory tract infections, suggesting that despite the distinct causative pathogens, the underlying response mechanisms in the blood draw on a common repertoire of immune programmes. This convergence helps explain why children with different infections can present with similar clinical pictures while still carrying pathogen-specific fingerprints detectable at the transcriptomic level.

The team also performed cellular decomposition analysis, a computational approach that infers the relative abundance of different immune cell types in whole blood from their gene expression patterns, avoiding the need for physical cell sorting. The results were sobering. Compared with healthy children, severe SARS-CoV-2 was associated with the depletion of 22 distinct cell populations, RSV-associated illness with 16, and pulmonary tuberculosis with 21. Depletion of circulating lymphocyte subsets is a recognised feature of severe respiratory viral infections and is thought to reflect both redistribution of cells to infected tissues and infection-induced cell death. The breadth of this depletion in severe paediatric COVID-19 underscores how profoundly the cellular architecture of the immune system is remodelled when disease becomes critical.

Perhaps the most clinically consequential finding is the set of genes capable of discriminating between disease states. The study identified transcripts that distinguish asymptomatic or mild SARS-CoV-2 infection from severe disease, that separate healthy children from those with severe COVID-19, and that differentiate RSV-associated illness and pulmonary tuberculosis from one another and from SARS-CoV-2. IFI27, an interferon-stimulated gene frequently implicated in antiviral responses, appears among the leading markers for both severe SARS-CoV-2 and RSV, while myeloid genes such as MMP8 and DEFA4 anchor the tubercular signature. The authors highlight these discriminators as potential future therapeutic targets, and they could also inform the development of diagnostic tests that distinguish bacterial from viral lower respiratory disease at the point of care, a longstanding goal in paediatric medicine where antibiotic stewardship is critical.

The work carries particular significance for low- and middle-income countries, where the study was conducted and where childhood pneumonia remains a leading cause of death. Most transcriptomic studies of COVID-19 have been performed in high-income settings with predominantly adult cohorts, leaving a gap in knowledge about immune responses in children from populations that bear a disproportionate burden of respiratory disease. By embedding this analysis within well-characterised South African birth and pneumonia cohorts, the researchers have generated reference data that reflect the epidemiological realities of the communities most affected. The study was approved by the research ethics committee of the University of Cape Town and the Western Cape Provincial Research Committee, with written informed consent obtained from mothers at enrolment and renewed annually, and secondary analysis approved by the University of Southampton.

As with any transcriptomic study, the findings describe associations between gene expression and disease state rather than proving causation, and whole blood sampling captures a systemic snapshot that may not fully mirror events within the lung itself. Nevertheless, the convergence of evidence, from differentially expressed genes through pathway enrichment, co-expression networks and cellular deconvolution, paints a coherent picture in which severe paediatric COVID-19 resembles the adult syndrome in its immunological architecture, while mild infection in children is marked by a far more restrained response. The gene signatures identified here, including OLFM4, IFI27, IGF2BP3 and their companions across disease groups, now provide a molecular framework for future studies of why some children deteriorate while most do not, and for the design of interventions aimed at tipping that balance.

Subject of Research: Immune gene expression differences in children with SARS-CoV-2 infection compared with other lower respiratory tract infections

Article Title: Immune transcriptomic differences in paediatric patients with SARS-CoV-2 compared to other lower respiratory tract infections

Article References: Kitaba, N. T., Workman, L., Cohen, C., Baralle, D., Kong, E., Botha, M., Johnson, M., Goldblatt, D., Nicol, M. P., Holloway, J. W., & Zar, H. J. (2026). Immune transcriptomic differences in paediatric patients with SARS-CoV-2 compared to other lower respiratory tract infections. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14343-x

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14343-x

Keywords: SARS-CoV-2, COVID-19, children, transcriptomics, gene expression, RSV, pulmonary tuberculosis, lower respiratory tract infection, neutrophil degranulation, interferon signalling, WGCNA, South Africa