Figure 4. Context‐dependent promoter memory and model.

1. Interval‐dependent regulation of promoter memory. Cells where treated with four consecutive Msn2 pulses (75% induction level) with 5‐min duration. The intervals between the pulses were 5, 7.5, 10, 15, and 20 min, respectively (left column). Population averages of the time‐varying transcription rates were calculated for DCS2 (middle column) and SIP18 mutant D6 (right column) considering all cells per condition (responding and non‐responding cells). Solid lines correspond to the mean of the population‐averaged transcription rate calculated over five independent inference runs and shaded areas mark two times the standard error above and below the mean.
2. Toy model of context‐dependent promoter manifestations. We considered a four‐state promoter model with complex, nonlinear Msn2‐dependent transition rates (top row). Green and red arrows indicate transitions, which are promoted or repressed by Msn2, respectively. Gray arrows correspond to Msn2‐independent transitions. We simulated the promoter response to all thirty Msn2 inputs and quantified its dynamics by calculating the expected total number of transitions between all states (middle row heatmaps; blue show transitions with high probability (e.g., state 2 is rarely occupied in the middle scenario (High Msn2))). Depending on the Msn2 inputs, certain state transitions are favored, while others are effectively repressed. Therefore, different classes of dynamical inputs can reveal distinct manifestations of the same promoter (bottom row).