Subtyping children with asthma by clustering analysis of mRNA expression data.

Abstract

Background: Asthma is a heterogeneous disease. There are several phenotypic classifications for childhood asthma. Methods: Unsupervised consensus cluster analysis was used to classify 36 children with persistent asthma from the GSE65204 dataset. The differentially expressed genes (DEGs) between different asthma subtypes were identified, and weighted gene co-expression network analysis (WGCNA) was carried out. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis was performed for DEGs and critical gene modules. Protein–protein interactions (PPI) were constructed to obtain the hub genes. Finally, differences in the immune microenvironment were analyzed between different subtypes. Results: Two subtypes (C1, C2) were identified using unsupervised consensus clustering. The DEGs between different asthma subtypes were mainly enriched in immune regulation and the release of inflammatory mediators. The important modular genes screened by WGCNA were mainly enriched in aspects of inflammatory mediator regulation. PPI analysis found 10 hub genes (DRC1, TTC25, DNALI1, DNAI1, DNAI2, PIH1D3, ARMC4, RSPH1, DNAAF3, and DNAH5), and ROC analysis demonstrated that 10 hub genes had a reliably ability to distinguish C1 from C2. And we observed differences between C1 and C2 in their immune microenvironment. Conclusion: Using the gene expression profiles of children’s nasal epithelium, we identified two asthma subtypes that have different gene expression patterns, biological characteristics, and immune microenvironments. This will provide a reference point for future childhood asthma typing and personalized therapy.