Fig. 2: Age-related changes in CD8+ T cell, CD4+ T cell, and B cell subsets.
From: Reading the immune clock: a machine learning model predicts mouse immune age from cellular patterns

a FlowSOM clustering identified six metaclusters in the t-SNE map. b Each metacluster was named using 12 signature markers for cell identification (CD8+ T cell: CD3ε+, CD8α+; CD4+ T cell: CD3ε+, CD4+; B cell: CD19+, B220+; cDC1: CD11c+, XCR1+, CD8α+; cDC2: CD11c+, CD11b+, CD172α+; Macrophage: Ly-6C+, F4/80+, CD11b+) The expression level of each marker was represented as a heatmap. c Age-related percentage change for each identified cell. Data were presented as box plots (n = 10 per month). d For CD8+ T cell, CD4+ T cell, and B cell subtype analysis, metaclusters of each cell were extracted from FlowSOM to create new opt-SNE maps. The molecular expression levels in each cell population were expressed as heatmaps. e CD8+ T cell opt-SNE is expressed as density dots. The regions of molecules strongly expressed at locations where cell counts change over months are indicated by individual color lines. f CD8+ cell subtypes were analyzed. Data were presented as raincloud plots (n = 10 per month). g CD4+ T cell opt-SNE is expressed as density dots. The regions of molecules strongly expressed at locations where cell counts change over months are indicated by individual color lines. h CD4+ cell subtypes were analyzed. Data were presented as raincloud plots (n = 10 per month). i Based on the SPADE analysis, we examined the proportion of subpopulation bubbles (I–VII) within each B cell population. j B cell subtypes were analyzed. Data were presented as box plots (n = 10 per month). Mass cytometry data were analyzed using Cytobank version 10.6 Cell frequencies were expressed as percentages based on 50,000 CD45⁺ cells. Data are shown as mean ±S.D (n = 10 per month). In box plots, the centre line represents the median, box limits represent the first and third quartiles, and whiskers extend to data points within 1.5× the interquartile range. The reported p-values were obtained from a one-way analysis of variance (ANOVA) and correspond to two-sided statistical tests. (*p < 0.05, **p < 0.01, ***p < 0.001, n.s: not significant).