Advanced Bayesian inference methods for simulation-based models in cell biology

Project overview

Understanding the parameters which govern the cell cycle and spread of cells has wide impacts: allowing practitioners to gain new insights, assess treatments and to make predictions, to name a few. Realistic models often are stochastic and computationally expensive to simulate. This project aims at developing efficient Bayesian simulation-based inference methods which are calibrated to real scientific applications in cell biology.

The simulation model emulates experiments with melanoma skin cells which use Fluorescent Ubiquitination-based Cell Cycle Indicator (FUCCI) technology to differentiate between phases of the cell cycle. Of interest in this application is the estimation of proliferation and motility rates through the separate phases of the cell cycle. Exploring Approximate Bayesian Computation (ABC) and Bayesian Synthetic Likelihood (BSL) methods, we aim to identify and compare these methods when applied to a real application in cell biology.


  • Estimation of Proliferation and Motility rates of melanoma skin cells.
  • Development of parameter estimation methods in Cell Biology applications.
  • Once limitations of parameter estimations methods in cell biology are understood, new and more efficient algorithms can be developed.

Project team