PNNL

Abstract

Recent experiments have established unambiguously that biological systems can have significant cell to cell variations in gene expression levels due to intrinsic or external noise even in isogenic populations. In this study we present a new approach to model gene regulatory networks that allows for fluctuations in the gene expression levels. The new algorithm is based on a probabilistic approach and it can be implemented using stochastic simulation methods. Since repression or induction of the genes and the biological variations among isogenic populations are simultaneously modeled, new algorithm represents the cellular systems in the gene regulatory network simulations in a more realistic manner. We have implemented the new algorithm in the NWGene simulation program and have tested it on a set of the synthetic gene network library that was recently formed by Guet and co-workers (Science, 296, 1466, 2002) using bioengineering techniques. The linear correlation coefficient between the computed and experimental results was larger than 0.9 showing a good degree of agreement. We have investigated the robustness of the new model by analyzing the sensitivity of the results to the model parameters and have also investigated if the model can successfully account for the cases in which chemical inducers are added to change the binding affinity of the transcription factor to the promoter sequences that it recognizes. These additional tests showed that the new algorithm is very successful in explaining the experimental data and that the model is robust and fairly insensitive to the model parameters.