Fig. 5: B-cell types are associated with distinct subsets of T-cells.

A B/T-cell cluster co-occurrence network by force-directed graphing as in Fig. 3B. Probabilistic model-based co-occurrence analysis (2-sided test) was performed using B/T-cell cluster information from 73 FL and 34 rLN samples without additional adjustment for multiple testing (see Methods). Each node represents a B- or T-cell PG cluster. Each B-cell cluster is colored separately; T-cell clusters are colored according to landmark group. Communities local to normal B-cells and tumor B-cell clusters B01 and B02 are shown; see Supplementary Data 10 for the full network image. B Cell abundance correlation plots. Each dot represents an individual sample. B- and T-cell cluster % values are among total B- and T-cells in each sample, respectively. Spearman rank correlation rho (R) and p-values (2-sided) are indicated. C Tfh cell content by tumor type. Each dot represents an individual sample; Tfh % values are among total T-cells within each sample. Tumor type A, n = 24 samples; type B, n = 10 samples; type NOS, n = 39 samples; type rLN, n = 34 samples. *p = 1.15e-02; **p = 6.20e-03 (A vs. B) and p = 5.18e-03 (A vs. rLN); ***, p = 1.75e-04; ****,p = 8.01e-07 (Kruskal-Wallis test with post-hoc Dunn test). EM effector memory, CM central memory, Tfh T follicular helper, Th1 T helper type 1, Treg T regulatory, MC metacluster, NOS not otherwise specified, rLN reactive lymph node.