Fig. 6. A facultative pausing model reproduces the live cell transcription data and predicts a long-lived pause.

A Schematics of the different models used to fit the live cell HIV-1 transcriptional data. Polymerases are represented by small orange balls. B Fits of the experimental survival functions. Graphs represent the survival functions reconstructed from the live cell data for the High Tat, Low Tat, and No Tat conditions, with the part deriving from the short and long movies in red and green, respectively. Blue line: fit of the 3-state model with a facultative pause; “+“: fit of the 3-state model with an obligatory pause; “x”: fit with a facultative pause. x-axis: time intervals between successive initiations events, in seconds and in log₁₀ scale. y-axis: probability of Δt > x (log₁₀ scale). C Model scores. The graph depicts the score of each model (inverse of the minimal value of the fitted Objective Function), for each of the model and cell line. D, E Pausing characteristics predicted by the models. D Predicted pausing times, for the relevant models and cell lines (see text for details). E Predicted pausing frequencies (in %), for the indicated cell line and model. For the model with the facultative pause and systematic abortion, the two indicated values come from the two branches of the model that could each correspond to the paused state (see the symmetric representation of the model M2 when k_release = 0 in Supplementary Notes Fig. S5). F, G Features of the model with the facultative pause. F The graphs represent the number of mRNA per cell measured by smFISH experiments (violet bars), or predicted from the model parameters (blue bars, with the center being the best fit value predicted from the model). Error bars are the standard deviation for the smFISH data (estimated from independent measurements; n= 3 for High Tat and Low Tat cells, and n = 4 for No Tat cells) and the confidence intervals for the prediction from the model (see Methods). Source data are provided as a Source Data file. G Estimated initiation rate (in s⁻¹), for the three cell lines (left), and the fraction of the cells with the promoter in the ON state (in %; right). The center is the best fit value predicted from the model and error bars are confidence intervals estimated during the fitting procedure (see Methods and Supplementary Note 3.3). Source data are provided as a Source Data file.