Skip to main content
. 2019 Dec 19;9:19471. doi: 10.1038/s41598-019-55852-x

Figure 6.

Figure 6

Differential abundances of lymphoid cell subsets in patients with early MS. (a) t-SNE map highlights mixed T and NK cell population (red dots). Heat map and cluster analysis of the mixed CD4+ T cell and NK cell population from all samples on the basis of the mean marker expressions (Panel B). Clusters are indicated by dendrograms. Heat colours show marker expression levels (red, high expression; dark blue, no expression). The bar graph shows mean cluster frequencies as % of parent (black number) and % of total cells (grey number). Cluster 3 and 6 (* in red) show differential abundances between healthy controls and patients with early MS. (b) The graph shows differences in frequency (%) of cluster 3 and 6 between the two studied groups (CON, grey; early MS, red). (c) Median expression levels as shown in the dot plot representation for randomly selected cells (n = 256/cluster) in the cluster 3 (orange), 6 (green) and all other clusters (blue). IL-6 was defined as a marker discriminating the differentially abundant cluster 6. (d) The representative t-SNE plot shows IL-6 expression levels and indicates the IL-6-expressing cell subset (orange circle, left image). The right image shows expression levels of IL-6 expressing cluster 6 (orange line), compared to the total cells (blue line). (e) t-SNE map highlights HLA-DR+ lymphocyte population (light blue dots). Heat map and cluster analysis of B cells from all samples on the basis of the mean marker expressions (Panel B). Cluster 7 (* in red) show differential abundances between the studied groups. (f) The graph shows differences in frequency (%) of cluster 7 between the two studied groups (CON, grey; early MS, red). (g) Median expression levels as shown in the dot plot representation for randomly selected cells (n = 256/cluster) in the cluster 7 (orange) and all other clusters (blue). MIPβ was identified as a marker discriminating the differentially abundant cluster 7. A P value < 0.05 at 10% FDR was considered statistically significant, determined using GLMM (**P < 0.01, unadjusted).