Figure 5: Transcription factor cooperativity in transcriptional reprogramming.

A. A linear model to decompose combinatorial TF perturbations from single TF data. The $\delta \mathrm{a}$ and ${\mathrm{c}}_2\delta b$ terms model independent contributions of individual TFs A and B, while the ${\mathrm{c}}^3\times \delta a\times \delta a$ term models cooperativity.

B. Distribution of $R^2$ across linear models across 801 TF pairs.

C. Across all 801 TF pairs, shown is the distribution of $c_1$ and $c_2$ values in the linear model.

D. Distribution of the number of cells used in each application of the linear model.

E. Visual embedding of TF pairs based on parameters of the linear model.

F. The bar chart indicates the number of TF pairs with the following relationship between constituent TFs: redundant, dominant, cooperative, neomorphic, and none.

G. Feature plots indicating TF pairs with redundant, dominant, cooperative, and neomorphic relationships between constituent TFs. Example TF pairs are indicated.

H. Example TF pairs with redundant (Hey2 + Nkx2.5), dominant (Hand2 + Wt1), and cooperative (Atf3 + Gata6) relationships. For each of the three plots, shown is the transcriptome expression pattern for: cells induced with each individual TF (top 2 rows), cells induced with both TFs (third row), and the prediction from the linear model (bottom row). The model parameters are indicated above each plot.

I. TF interaction relationship in the context of epicardial reprogramming.

J. For the epicardial reprogramming batch, cells were ranked by pseudotime and colored by whether they contain the indicated TF pairs.

K. qPCR analysis of Gpm6a and Upk1b expression for pairwise and individual TFs.