In living cells, molecular noise originates from thermal fluctuations and noisy processes such as transcription and translation. Hence, random fluctuations affect concentration levels of mRNA and protein, whose expression can be viewed as a stochastic process.
The model of molecular noise that we have developed uses Langevin equations. We have implemented this model in GeneNetWeaver (GNW) to generate data sets for assessing the performance of network inference methods in the fourth and fifth editions of the DREAM Project, an annual community-wide network inference challenge.
The integration of the ODE model of gene regulation in GNW leads to noiseless gene expression. If molecular noise is selected, the ODE model is replaced by a system of stochastic differential equations (SDEs) called Langevin equations which is described here.
The measurement noise depends on the technology used to monitor gene expression concentrations and is modeled independently of the molecular noise. GNW implements Gaussian and log-normal models of experimental noise as well as a model of noise observed in microarrays.