Figure 4. Elongation rate explains transcriptional stability and the steady-state output.

(A) Models of Pol II elongation (left), and the expected steady-state GRO-seq density (D) and the inverse elongation rates (v⁻¹) for each model (right). The termination model proposes that the decrease in D is a combination of Pol II termination and increasing elongation rate, while the acceleration proposes that the decrease in D is primarily a consequence of increasing elongation rate. (B) The steady-state GRO-seq density (D) and inverse elongation rates (v⁻¹) from average transition points of the FP measurements. (C) Stages of transcription determining the mRNA level following the productive elongation stage. (D) Correlation plot between GRO-seq gene body density as a measure of nascent transcription and RNA-seq as a measure of mRNA steady state level. (E) Non-linear correlation between nascent transcription level and mRNA level in highly transcribed genes. To determined the monomial degree of the correlation, a LOESS fit was used for the scatterplot in D, and the slopes of the LOESS fit in the higher 50 percentile and the lower 50 percentile were derived. (F) Elongation rate correlates with GRO-seq density. Correlation plot was determined from the z-scores of the elongation rates and the gene body GRO-seq densities. (G) Correlation of mRNA steady state level and GRO-seq of the 938 mid elongation rate genes (12.5–25 min), and the monomial degree of the correlation derived from the slope. (H) Correlation plot of the mRNA production rate (GRO-seq density multiplied by the elongation rate) and mRNA steady state level of the same genes, and the monomial degree of the correlation.

DOI: http://dx.doi.org/10.7554/eLife.02407.017

Figure 4—figure supplement 1. Monte-Carlo modeling of the acceleration and the termination hypotheses.

(A) Simulation time-course of change in Pol II distribution after a block of Pol II entry into the gene body assuming an intrinsic acceleration model (top) and a termination model (bottom). See supplementary methods for the details of the simulation. (B) Pol II density D and inverse elongation rate v⁻¹ plot of a simulation assuming a termination model as explained in the top panel of Figure 4A. (C) Pol II density D and inverse elongation rate v⁻¹ plot of a simulation assuming a mixed rates model. In this model, slower and faster populations of Pol II co-exists, and slower Pol II terminates more frequently leaving the faster Pol II more abundant in more downstream regions. (D) Pol II density D and inverse elongation rate v⁻¹ plot of a simulation assuming a intrinsic acceleration model as explained in the bottom panel of Figure 4A. (E) Scatterplot of the slope of v and the slope of D⁻¹ for multiple acceleration models and termination models. See supplementary methods for the details of the simulation.

DOI: http://dx.doi.org/10.7554/eLife.02407.018