Fig. 7.

Alternative model settings. a Upon influenza infection, termination is altered genome wide, thus affecting observed mRNA counts; μ_X,uninf. (μ_X,inf.) and ν_uninf. (ν_inf.) are the mean expression and noise levels in uninfected (infected) cells (measured as the distance from solid-line prediction as in Fig. 5), respectively, computed from the scRNA-seq data of [79]. The majority of genes (58%) reports an increase in noise and independently a decrease in mean expression upon infection (upper-left quadrant in scatter plot is the most populated). b Noise plots of mRNA abundances simulated according to the model of Fig. 6 with same parameter values except l and d as in legend; incrementing the mRNA degradation rate d suppresses μ_X for both values of l, while increasing ${\text{CV}}_X^2$. c Noise plots of mRNA abundances in a variant of the looping model; recycling is allowed but PolIIs cannot pile up before initiation, with noise being virtually unaffected when l is tuned from 0.95 to 0; other model parameters are as in Fig. 6b except (γ,β)=(50,100). The settings of b and c cannot explain the noise patterns observed in HBB and env transgenes