François Lamoline (University of Luxembourg, Luxembourg)

Title: Gene expression modelling from cell population snapshot data using optimal mass transport

Modelling gene expression is a central problem in systems biology. Accurate predictive models provide powerful tools for understanding cellular mechanisms and exploring the regulatory relations between genes. Perturbations of these regulatory structures affect the cellular functions. The ability to predict the effects of these perturbations is critical for finding sources of complex diseases and developing new treatments. Recently, single-cell techniques have enabled to perform sequencing at the level of individual cells for a large number of cells at a time. Unfortunately, the cells are destroyed in the measurement process, and so the data consist of population snapshots at different times. Traditional methods aim at modelling from time series data and cannot utilise the full information in the richer single-cell data. Therefore, these new sequencing techniques have raised the need of tailored computational methods for modelling the gene expression from single-cell data. In this talk we introduce new methods based on the 18^th century problem of optimal mass transport. The idea consists in tracking the evolution of the distribution of cells over time and finding the dynamical system that minimises the transport cost between consecutive time points. The performance of the methods is compared in numerical experiments.

Link to the seminar here

People outside Université de Namur with a Microsoft Teams account may join. For any questions, write to riccardo.muolo@unamur.be