Fig. 5: Prediction of accelerated immune-aging due to obesity.

a The t-SNE map expresses the number of cells as contour lines. Depending on the aging and obesity conditions, various cell changes appear at the same location on the map (5N, 17N; n = 10, 5H, 17H; n = 9). b Population changes of six cell populations (CD8⁺ T cell, CD4⁺ T cell, B cell, cDC1, cDC2, macrophage) separated using FlowSOM (n = 10 per month). Data were shown as box plots. Although both c, d represent predicted immune ages, c provides a statistical summary of group distributions, whereas d highlights heterogeneity at the level of individual samples (that may not be as easily appreciated in c). c Immunological age distribution by group. d Heatmap of actual vs. predicted immune age (per sample). e 2D PCA comparison with 95% confidence ellipses. f Correlation matrix of the top 10 PC2-contributing immune markers derived from PCA including all groups (2M–20M and 5N/5H/17N/17H). These markers represent shared molecular drivers of PC2 variation, reflecting combined contributions of aging and obesity, and highlight common pathways underlying immune age shifts. Cell percentages were calculated relative to CD45⁺ cells 50,000. Mass cytometry data were analyzed using Cytobank version 10.6. Data are shown as mean ±S.D. 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).