Figure 6. Single-Cell RNA Sequencing of Thoracic SCTNs in Hoxc9 Mutants.

(A) t-SNE visualization of putative joint and separate cell clusters using all cells from both control and Hoxc9 mutants. Cells were clustered in three sets: (1) control only (as represented in Figure 3), which are labeled as 1–8 on the plot, corresponding to SCT1 through SCT8, (2) mutant-only, which are labeled as m1, m2, m3, m6, and m8 on the plot, and (3) both control and mutant cells; these joint clusters are color-coded on the plot. In general, the joint clusters agree with the independent clustering of control-only and mutant-only cells and suggest the correspondence across the two sets of cells. For example, joint cluster 1 (blue) contains cells mostly from control SCT1 (1) and mutant cluster 1 (m1), while joint cluster 3 (brown) contains cells mostly from control SCT3 (3) and mutant cluster 3 (m3).

(B) As in Figure 3C, dot plot representing the number of cells in each cluster originating from each control and mutant dissection. For the mutant, each cluster was assigned to its corresponding control SCT cluster, based on the joint clustering shown in (A). The numbers for the control dissections are the same as in Figure 3C.

(C) Alternative approach to assign Hoxc9 mutant cells to control clusters. The heatmap shows the classification probabilities for each mutant cell (row) using a random forest classifier trained on the eight control cluster identities (columns). The colorbar on the left indicates the mutant cell cluster identity (m1, m2, m3, m6, and m8). The overall classification closely resembles the result from the joint clustering shown in (A); for example, cells from m1 have high classification scores for control SCT1, whereas cells from m3 tend to be most strongly assigned to SCT3.

(D) As in Figure 3D, barplot showing expression (TPM) values for selected genes in the cell clusters derived from Hoxc9 mutant SCTNs.

See also Figure S6.