Calibration of a stochastic model of gene expression including feedback and extrinsic noise sources

Summary form only given. In recent years, fluorescence microscopy and constructed reporter strains have been applied to investigate the stochastic behavior of gene expression in single cells. While many of these studies have focused on the magnitude of the fluctuations as measured by variance, a great deal of additional information concerning the timing of the fluctuations is contained in the autocorrelation function. Here we calibrate a stochastic model of gene expression that explicitly accounts for the effects of intrinsic and extrinsic noise using comparisons of the experimental and simulated autocorrelation function as basis of the goodness of fit. The experimental data set includes unregulated and negative auto regulation cases of similar gene circuits. The calibration process is computationally demanding because it requires many evaluations of the stochastic simulation model. We demonstrate how a combination of computational techniques can be employed to make the problem tractable. The calibrated model demonstrates how feedback affects the manner in which extrinsic noise sources at the transcription and translation level are propagated through the circuit