 |
Center for Nonlinear Studies |
Despite vast amounts of biochemical information, it remains difficult to understand or predict the quantitative responses of signal transduction and gene regulation pathways. In this presentation, I discuss new approaches to integrate dynamic single-cell and single-molecule experiments with discrete stochastic analyses. I use these methods to identify models capable of making quantitative predictions for transcriptional dynamics on the level of single cells. I illustrate the power of this approach in a combined experimental/computational investigation of the osmotic stress response pathway in Saccharomyces cerevisiae. After generating several thousand different model structures, we use simple parameter estimation and cross-validation analyses to exclude models that are either too simple or too complex to be supported with the available data. Through a process of iterative experiment design, we eventually select a single quantitative model with the greatest predictive capability. This model yields insight into several dynamical features, including multi-step regulation and low-pass filtering. Furthermore, the model predicts the transcriptional dynamics of cells in response to new environmental and genetic perturbations. Since our approach is general, it can facilitate a predictive understanding for signal-activated transcription in any gene, pathway or organism.
Host: Peter Loxley, loxley@lanl.gov