Inferring graphs of interactions between genes has become a textbook case for high-dimensional statistics, while models describing gene expression at the molecular level have come into their own with the advent of single-cell data. Linking these two approaches seems crucial today, but the dialogue is far from obvious: statistical models often suffer from a lack of biological interpretability, and mechanistic models are known to be difficult to calibrate from real data.

We recently introduced two strategies that exploit time-stamped scRNA-seq data, where single-cell profiling is performed at successive time points after a stimulus: a mechanistic network model driven by transcriptional bursting, and a scalable inference method seen as model calibration. Thanks to this correspondence, I will show how the two approaches can be combined to obtain both a plausible network and a quantitative simulation tool. The edges of the network ultimately have an explicit causal definition in a probabilistic paradigm, and the strong variability observed in scRNA-seq data can in fact be explained mainly by biological stochasticity, which then plays a functional role.

This work is in collaboration with Elias Ventre, Thibault Espinasse, Gérard Benoit and Olivier Gandrillon.