Figures and data

Comparison between the exact Gillespie algorithm and the DGA for simulating chemical kinetics. (a) Example of kinetics with N = 3 reactions with rates ri(i = 1, 2, 3). (b) Illustration of the DGA’s approximations: replacing the non-differentiable Heaviside and Kronecker delta functions with smooth sigmoid and Gaussian functions, respectively. (c) Flow chart comparing exact and differentiable Gillespie simulations.

Flowchart of the parameter optimization process using the DGA. The process begins by initializing the parameters

Two-state gene regulation architecture. (a) Schematic of gene regulatory circuit for transcriptional repression. RNA polymerase (RNAP) binds to the promoter region to initiate transcription at a rate r, leading to the synthesis of mRNA molecules (red curvy lines). mRNA is degraded at a rate γ. A repressor protein can bind to the operator site, with association and dissociation rates

Accuracy of the DGA in simulating the two-state promoter architecture in Fig. 3(a). Comparison between the DGA and exact simulations for (a) steady-state mRNA distribution, (b) moments of the steady-state mRNA distribution, and (d) the probability for the promoter to be in the “ON” or “OFF” state. (c) Ratio of the Jensen-Shannon divergence JSD

Gradient-based learning via DGA is applied to the synthetic data for the gene expression model in Fig. 3(a). Parameters

Gradient-based learning via DGA is applied to the synthetic data for the gene expression model in Fig. 3(a). Parameters

Fitting of experimental data from Ref. [37] using the DGA. (a) Comparison between theoretical predictions from the DGA (solid curves) and experimental values of mean and the Fano factor for the steady-state mRNA levels are represented by square markers, along with the error bars, for two different promoters, lacUD5 and 5DL1. Solid curves are generated by using DGA to estimate

Design of the four-state promoter architecture using the DGA. (a) Schematic of four-state promoter model. (b) Target input-output relationships (solid curves) and learned input-output relationships (blue dots) between activator concentration [c] and average mRNA production rate. (c) Parameters learned by DGA for the two responses in (b). (d) The sharpness of the response

In panels (a) and (b), we plot the ratio of the Jensen-Shannon divergence JSD

Error bars estimation for asymmetric loss function.