Fig. 1

COVID-19 patients, especially those with severe infection, showed increased levels of regulatory molecules and decreased levels of multiple cytokines in peripheral blood T cells. a Heat maps comparing peripheral blood leukocyte subset concentrations in healthy (n = 6), mild (n = 10), and severe (n = 6) patients. Rainbow-colored squares represent mean values of each group. Red-black-green squares represent log 10 P values, and white asterisk indicates P ≤ 0.05 by post hoc ANOVA test. H, healthy; M, mild; S, severe. b Comparisons of IL-6, TNF-α, and sCD14 plasma concentrations in healthy, mild, and severe groups. n.s., P > 0.05, *, P ≤ 0.05, by Kruskal–Wallis test. c Comparisons of expression levels of activation-, regulation-, and function-related molecules in CD4+ and CD8+ T cells among groups. Rainbow-colored squares represent mean positive cell rate for each group. d Comparisons of cell expression modules of exhaustion-related (CTLA-4, PD-1, and TIGIT) and function-related (IFN-γ, TNFα, and IL-2) molecules in CD4+ or CD8+ T cells among groups. “Single” indicates that cell only expresses one of the three molecules, “Multi” indicates that cell expresses at least two of the three molecules, “Non” indicates that cell expresses none of the three molecules. Red-yellow-blue squares indicate average cell expression rates of different modules of three groups, respectively. e Correlation network analysis of markers with significant differences among groups. Nodes are colored based on cell type for three groups. Node size indicates relative strength value according to centrality analysis. Thicker lines indicate more correlated genes. Green lines represent significantly positive Spearman’s correlation coefficients ≥0.40; red lines represent significantly negative Spearman’s correlation coefficients ≤−0.40. f Hierarchical clustering of participants based on all immunological risk indicators