We have developed a new model of transcriptional control. A central goal of this model is to show how modular enhancers arise from groups of binding sites, and more generally to predict the expression patterns of modular enhancers from knowledge of their binding sites.
The model has a three-layer organization. The first layer describes the binding of activators (A) and repressors (Q) to the regulatory region of a gene, and incorporates the effects of repression by competition and quenching. The second layer describes the adapter molecules binding to DNA-bound activators, and incorporates the effect of direct repression. Finally, the activation of transcription is modeled by an Arrhenius mechanisms in which activating adapters lower the activation energy barrier.
Currently, we are modeling the regulation of the Drosophila even-skipped gene. We measure the concentration of lacZ mRNA driven by different eve regulatory regions under the control of the endogenous eve promoter (see Quantitative Data Acquisition). These data are combined with our existing dataset from the (FlyEX) database which contains concentrations of the transcription factors that are known to directly bind the eve regulatory sequences. The model will then be fit to this expanded dataset by large-scale numerical optimization. If one set of model parameter values enables the model to fit the data, this would suggest that the model incorporates the most important mechanisms of transcriptional control. Analysis of these parameter values (such as the binding affinities, the quenching strengths and activation efficiency) can then help us in understanding and predicting transcriptional regulation.