Research

Mathematical and statistical modelling

We develop mechanistic and statistical models at different scales: from intracellular signalling to tumours and human populations. Models include differential equation models, agent-based models and survival models.

Inference and identifiability

With the growing availability of data, integrating models with empirical observations has become a key part of cancer research. We develop methods to infer model parameters from data, and study when such inference is (and is not) possible.

Uncertainty quantification

Model-based predictions are only useful if we know how much to trust them. We quantify uncertainty in model parameters and the resulting dynamics. We also develop diagnostics for Markov chain Monte Carlo.

Equation learning

Machine learning methods can be used to learn the equations that govern dynamical systems from data. We use physics-informed neural networks to learn partial differential equations from cancer cell data, and propose pipelines for selecting hyperparameters in such applications.