Fig. 3

FlowSOM minimal spanning tree (MST) clustering identifies eight novel γδ T -cell subsets. Total γδ T cell flow data from middle aged to elderly healthy donors were clustered into 100 nodes with subsequent automated meta-clustering into eight γδ T cells subsets (meta-clusters) using Vδ1, CD161, CD45RA and, CD27, CD8 and CD4 clustering channels and the FlowSOM algorithm. The relationships between the nodes (which are most like each other) are shown by the spanning tree with similar nodes placed close together on the plot. Expression intensity of each clustering marker on the same spanning tree plot is shown in the six diagrams (A) Vδ1, (B) CD161, (C) CD45RA, (D) HLA-DR (E) CD27, (F) CD8 and (G) CD4. Dark red represents maximum expression, and dark blue represents no expression of the given marker as shown in the colour bar. A coloured halo around the node indicates the meta-cluster to which each node belongs. Expression levels of each marker in each cell meta-cluster is summarized in Table 1