The current clinical classification of coronavirus disease 2019 (COVID-19) does not adequately capture the biological heterogeneity observed among patients. To address this gap, the present study aimed to identify distinct subtypes of severe COVID-19 through unsupervised clustering analysis. We analyzed nine publicly available RNA sequencing datasets of peripheral blood samples from the GEO database. After identifying differentially expressed genes (DEGs), we applied a consensus clustering algorithm to classify the samples into distinct subtypes. To further characterize these subtypes, we performed gene set enrichment analysis and assessed immune cell infiltration to understand their underlying biological mechanisms. Based on the 139 upregulated DEGs of severe COVID-19 infection, patients were divided into subtype A, subtype B, and subtype C, each with different molecular and cellular characteristics. Subtype A was characterized by activated neutrophils that undergo degranulation and respond to bacteria or fungi. Subtype B showed significant activation in canonical pathways associated with interferon-alpha/beta signaling. Subtype C was characterized by immune cell activation associated with pathways of mitotic and cell cycle. These results facilitate the development of a precise classification framework, which informs the design of molecular diagnosis and provides actionable guidelines for stratified therapy in severe COVID-19 infection in the future.
Qiao et al. (Fri,) studied this question.