Figure 4

Microarray transcriptional comparison of FACS sorted cells. Microarray transcriptional comparison of FACS sorted thymic DP^hi CD24^hi NK1.1⁻ CD1d-tet⁻TCRβ⁺ (Population (P) 1 (P1)) cells, DP^hi CD24^hi CD44^lo NK1.1⁻ CD1d-tet⁺TCRβ⁺ (P2) cells, CD4⁺CD8⁻CD24^hi NK1.1⁻CD1d-tet⁺TCRβ⁺ (P3) cells and CD4⁻CD8⁻ CD24⁻ NK1.1⁺CD1d-tet⁺TCRβ⁺ (P4) cells from individual NOD.Va14^tg /1 mice (n = 7). (A) Gating strategy for sorting of populations. (B) Principal component (PC) analysis of transcript expression. The three largest PC explain > 58% of variation; samples from each of the four cell types cluster together into groups, and the groups are distributed separately across PC1 (X axis), which explains > 29% of variation. A graphic illustrating the proposed NKT cell maturation pathway is provided for comparison. (C) Undirected, weighted transcriptional regulatory network generated from thymic NKT cell subsets P1–P4 (n = 7/population), consisting of 1,929 transcripts (nodes) and 10,626 pair-wise correlations (edges) across twelve co-regulated clusters (modules), named: Black (587 nodes), Cyan (71), GreenYellow (177), Grey (32), LightCyan (42), LightGreen (26), Magenta (127), MidnightBlue (68), Pink (525), Purple (188), Salmon (37) and Tan (49). (D) Heat map encoding of student’s t-test p values generated by pair-wise comparisons of transcript abundance across Transition 1 (P1 to P2) mapped onto the transcriptional regulatory network illustrated in panel C. (The WGCN depicted in (C,D) were visualised using a heat map in Cytoscape 3.8.0; https://github.com/cytoscape/).