Skip to main content
. 2020 Mar 17;16(3):e9174. doi: 10.15252/msb.20199174

Figure 1. Inferring direct regulation using many TF induction experiments.

Figure 1

  1. In each experiment, one transcriptional regulator with an inducible promoter is rapidly overexpressed in response to 1 μM β‐estradiol.
  2. Example of three genes (labeled B, C, and D) responding with different kinetics following induction of regulator A.
  3. Hypothetical example of a regulatory cascade in which an induced transcriptional regulator A directly inhibits C and directly activates B. B, in turn, directly activates D.
  4. In practice, we do not know that A regulates D via B and instead want to infer such regulatory relationships. In this example, direct regulation of D by B is only one hypothesis that is consistent with the data—all viable hypotheses are shown by dashed lines. A could directly activate D, C could inhibit D, or A could regulate an unmeasured confounder U which is the true regulator. Direct regulation by a variable Y, which is independent of A, is not possible since the timecourse begins at steady state.
  5. Integrating the A induction timecourse with a second induction timecourse, which perturbs B without perturbing A or C, allows us to narrow down D's possible sources of regulation. In this case, U may still be a possibility if it is remains correlated with B.
  6. Overview of data and analysis performed in this study. Over 200 induction experiments were constructed allowing for many opportunities to resolve ambiguous regulation.