Using Stochastic Models to Interpret, Predict and Design Single-cell Experiments


Modern experimental advances have made it possible to measure DNA, RNA, and protein dynamics at the level of single molecules in single cells. These discrete data are often noisy, but they can be highly informative when combined with advanced computational analyses. I will present our recent advances to integrate discrete stochastic models with different single-cell assays. I will introduce modern experiments that are used to quantify mRNA and protein dynamics at single-molecule resolution. Next, we will derive strict bounds on the likelihood that observed single-cell data come from hypothesized models, and we will show how these bounds can be used to reduce computational efforts with no sacrifice to accuracy or precision. Then we will explore new computational approaches to design optimal single-cell experiments that maximize the expected information contained within an experiment. We will illustrate the integration of these experimental and computational tools to build predictive models for mechanistic biochemical processes in yeast and human cells, including spatiotemporal cell-signaling, single-gene transcription activation, nascent mRNA elongation, mRNA transport, mRNA-translation activation, sequence-dependent nascent protein elongation, and viral frameshifting.