The study of dynamical systems has long focused on the characterization of their asymptotic dynamics such as fixed points, limit cycles and other types of attractors and how these invariant sets change their properties as systems parameters are varied. More recently, however, the importance of transient dynamics, especially of long transients and sequential transitions between them, has been increasingly recognized in various fields including ecology, neuroscience and cell biology. Among several possible origins of long transients, ghost attractors have received particular attention due to interesting dynamical properties in non-autonomous systems, new theoretical developments, and an increasing number of systems that empirically show dynamics consistent with ghost attractors. In this talk, I will present a new formal definition for ghost attractors of fixed points which allows us to not only develop an algorithm for finding ghosts (available as open-source python package: PyGhostID) but also to identify novel types of ghosts with unique properties, leading to surprising phenomena such as bifurcations involving ghosts. I will further present how our framework can be used to gain new insights into the transient dynamics of a wide range of systems, ranging from information processing and decision making in neuronal circuits, tipping of coupled climate elements and potentially also tissue regeneration following a heart attack.

Fatty acid synthase (FASN)-catalyzed endogenous lipogenesis is a hallmark of cancer metabolism. However, whether FASN is an intrinsic mechanism of tumor cell defense against T cell immunity remains unexplored. To test this hypothesis, here we combined bioinformatic analysis of the FASN-related immune cell landscape, real-time assessment of cell-based immunotherapy efficacy in CRISPR/Cas9-based FASN gene knockout (FASN KO) cell models, and mathematical and mechanistic evaluation of FASN-driven immunoresistance. Mathematical modelling shows that cancer cells engineered to carry a loss-of-function mutation in FASN exhibit an enhanced cytolytic response and an accelerated extinction kinetics upon interaction with cytokine-activated T cells. We also explore the effects of knock-out of other metabolic enzymes.

Volume-exclusion interactions play a crucial role in the collective behaviour and self-organisation of anisotropic particle systems, from liquid crystals to biological populations such as dense suspensions of myxobacteria and migrating cells. Even when particles have negligible volume, their mutual exclusion can generate a non-trivial effective volume in configuration space, strongly influencing macroscopic dynamics and the emergence of orientational order. In this talk, we first consider hard Brownian needles in two dimensions and use matched asymptotic expansions to derive a nonlinear kinetic equation for their density in position and orientation. We explore the resulting excluded-volume effects, including the emergence of an isotropic-to-nematic transition at increasing densities. We then extend the framework to finite-width rod-like particles, investigating how particle geometry and aspect ratio modify the effective interactions, and finally discuss active elongated particles, where steric interactions combine with self-propulsion to produce richer collective dynamics.

Cells make decisions across developmental biology, immunology, and synthetic biology. Yet, while the timing of these decisions can be as important as their outcome, mathematical models often focus primarily on stable cell states. In this talk, I will present a mathematical framework based on orbital equivalence to identify molecular mechanisms that can control the speed of cellular decisions while preserving their underlying dynamics. I will first address developmental tempo, where embryos from different species execute remarkably similar genetic programmes at different speeds. I will then turn to gene regulatory networks, showing how critical slowing down and transient dynamics can regulate decision times, and how analysing different network topologies can reveal design principles for controlling dynamical speed in synthetic and natural systems.

We develop mechanistic models at the cell and tissue level based on PhysiCell and PhysiBoSS, combining probabilistic Boolean networks for intracellular signalling with agent-based rules for cell growth, migration and interactions. These simulations have been applied to predict drug synergies, modes of metastasis, and responses to metabolic conditions. However, building patient-specific digital twins remains complex and cumbersome. Mechanistic models offer interpretability and causal reasoning but struggle with parameter identifiability and sparse patient-specific data. 

To address these problems we are working in embedding mechanistic agent-based simulations within an AI-assisted architecture. The system combines our multiscale simulator with a data integrator that learns patient-specific network structures and model parameters through a bidirectional co-design loop, enabling AI to inform mechanistic parameterisation while simulations generate training constraints for AI models. The initial developments are explored in the context ofthe simulation of perturbations in organoids with drugs and drug combinations and cancer progression and treatment response. 

In my view, this approach illustrates a paradigm for AI for Science where machine learning and mechanistic modelling work as integrated components of a unified framework for understanding and predicting complex biological systems.

Computational modeling of developmental processes and morphogenesis is challenging due to the need for cross-comparison between spatial features in experimental data and modeling outputs. In this talk, I will discuss my lab's efforts to derive quantitative metrics that can be used for parameter optimization and model evaluation of stem cell differentiation in 2D and 3D contexts and our resulting simulations and insights using off-lattice agent-based models. Furthermore, I will discuss our novel tools for near-single cell spatial characterization of metabolic processes that correlate, and potentially drive, differentiation.

Nature has survived billions of years of upheaval—yet coral reefs are collapsing, the Amazon is losing resilience, and lakes can flip into degraded states almost overnight. How can systems this battle-tested also be this brittle? In this talk, I'll argue that the answer lies in a central paradox of complex systems: the "Robust-Yet-Fragile" principle. The very features that confer resilience can also accumulate hidden vulnerabilities to novel or extreme shocks.

I'll walk through a series of empirical examples to show this duality in action, then formalize it through the lens of control theory's "conservation of fragility". The implication is unsettling: managing nature for resilience may be exactly what makes it fragile, and averting collapse requires learning to spot vulnerabilities hiding in plain sight.