New asthma expression models expand gene discovery in African-ancestry cohorts
Asthma genetics has long been shaped by uneven representation in reference datasets, and the gap is especially visible in studies of genetically regulated expression. In African-ancestry populations, where asthma burden remains high, the lack of tissue-specific prediction models has limited transcriptome-wide association studies and narrowed the search for biologically relevant risk genes. This analysis addressed that problem directly by building prediction resources for two asthma-relevant tissues: nasal epithelium and CD4+ T cells.
The study paired those new models with existing public prediction databases and applied them to a large African-ancestry cohort of 9,284 individuals. The goal was straightforward and ambitious: identify candidate asthma genes whose expression appears to be genetically regulated in tissue types connected to airway inflammation and immune signaling. The result was a broader and more ancestry-representative map of asthma susceptibility, with several genes emerging that were not captured by existing public resources.
Tissue-specific prediction improved the view of asthma biology
Investigators developed and validated gene expression prediction databases for unstimulated CD4+ T cells and nasal epithelial tissue using an elastic net framework. The new databases contained 8,351 genes for CD4+ T cells and 10,296 genes for nasal epithelium. Those are not abstract technical gains. In practical terms, they create a better bridge between genotype and disease biology in tissues that are directly involved in airway immune surveillance and epithelial barrier function.
A particularly important finding was the inclusion of four asthma loci — SCGB1A1, MUC5AC, ZNF366, and LTC4S — that could not be predicted with existing public databases. In asthma, that matters because airway epithelial programs and inflammatory mediator pathways often sit at the center of symptom generation, exacerbation risk, and chronic remodeling. If a tissue model cannot estimate expression for a relevant locus, the association signal may disappear before it can be linked to disease.
Prediction performance held up well against established databases and was most accurate in populations that shared ancestry with the training set, including African ancestry. This finding fits a familiar genetic principle: reference panels work best when the training population resembles the study population. When that condition is met, the probability of recovering meaningful expression-trait links rises, especially for genes whose regulation varies across ancestry backgrounds.
TWAS identified 17 candidate asthma genes in African-ancestry participants
Using the new and existing databases together, the transcriptome-wide association study identified 17 candidate causal asthma genes at an adjusted P value threshold of .1. Some signals were tissue-specific, while others appeared across tissues, suggesting that asthma risk in African-ancestry populations reflects both local airway biology and broader immune regulation.
One of the clearest tissue-specific signals involved IL33 in nasal epithelium. IL-33 is biologically plausible in asthma because it participates in epithelial alarm signaling and can amplify downstream type 2 inflammatory responses. When airway epithelial cells alter IL33 expression, the resulting immune cascade can affect eosinophilic inflammation, bronchial hyperresponsiveness, and exacerbation susceptibility. The nasal epithelium is not the lower airway, but it remains a relevant respiratory surface for studying upstream inflammatory programming.
Cross-tissue signals were also identified for CCNC and FBXW7. The fact that these genes emerged across tissue contexts suggests a more general regulatory influence rather than a single isolated epithelial or immune-cell effect. In a disease like asthma, where inflammatory signaling, epithelial injury, and immune-cell activation overlap continuously, cross-tissue associations can be especially informative because they point to pathways that may influence multiple steps in disease expression.
A single-sentence clinical takeaway stands out here: ancestry-matched expression models recovered asthma genes that older public resources missed.
Why ancestry-specific reference data changes the genetic signal
The study reinforces a basic but often underused concept in translational genomics: gene regulation is not uniform across populations, and reference mismatch can mute otherwise important signals. Existing public databases are useful, but they do not fully represent African-ancestry regulation patterns in asthma-related tissues. By adding CD4+ T cell and nasal epithelium prediction models derived with better ancestry alignment, the investigators improved the odds of capturing biologically credible gene-trait associations.
This is especially relevant in asthma, where the disease burden in African-ancestry populations is influenced by a blend of inherited susceptibility, environmental exposures, access to care, and inflammatory endotype variation. Genetically regulated expression does not explain all of that complexity, but it can identify genes that sit upstream of disease biology and are less visible in standard genome-wide association approaches alone. In that sense, TWAS can act as a functional filter, narrowing a broad genetic signal toward tissues and pathways that may be more directly actionable.
The appearance of SCGB1A1, MUC5AC, ZNF366, and LTC4S in the new databases also deserves attention because these loci align with known asthma-relevant biology. Airway secretory function, mucin production, transcriptional control, and leukotriene signaling are all plausible contributors to phenotype severity and symptom persistence. The study did not claim mechanistic proof for each association, but the tissue-specific predictions make the pattern biologically coherent rather than arbitrary.
A larger resource for future asthma gene mapping
The investigators concluded that adding these CD4+ T cell and nasal epithelium prediction databases to the public sphere should improve ancestry representation and strengthen power for future TWAS efforts. That is a concrete contribution, not a theoretical one. Better reference panels can increase discovery yield, reduce missed signals, and help explain why some disease associations have been difficult to replicate across populations.
The work also points toward a more careful model for asthma genetics research. Tissue selection matters because asthma is not driven by a single compartment. Blood-only approaches may miss airway-specific regulation, while airway-only models can overlook immune-cell contributions. The new resources try to cover both sides of that biology, which is why the combination of nasal epithelium and CD4+ T cells is clinically meaningful.
Even with those gains, the findings remain association-based rather than causal in the experimental sense. The identified genes warrant follow-up in functional studies that can test how altered expression changes epithelial responses, immune activation, and asthma phenotypes in African-ancestry populations. For now, the study establishes a stronger analytical foundation and makes one point unmistakable: without ancestry-representative tissue models, important asthma genes stay hidden.