Journées communes

• Pete Czuppon (I2M)
• Giada Fiandaca (INRIA – COMPO)
• Oksana Grente (BioSP – INRAE)
• Baptiste Oger (BioSP – INRAE)

Organisation de la journée :

• 10h : accueil café
• 10h30 – 11h30 : Giada Fiandaca

Title: Towards a Mechanistic Understanding of Tumour Phenotypic Heterogeneity Across Scales

Abstract: Tumour phenotypic heterogeneity is a major contributor to treatment failure and disease recurrence. It manifests across biological scales, from differences in cellular states and responses within clonal populations to changes in population composition driven by cellular plasticity and reciprocal interactions with the microenvironment. Continuous phenotype-structured population models offer a framework for describing this diversity, but phenotypic states and transitions often remain phenomenological. Recent advances in single-cell transcriptomics and live-cell imaging provide complementary molecular and dynamical data, opening opportunities to ground these descriptions in intracellular regulation.mIn this talk, I will present two ongoing projects that motivate this research direction. The first investigates heterogeneous responses to TRAIL, a signalling protein that can trigger programmed cell death, in clonal cancer cell populations. Combining longitudinal single-cell measurements with mechanistic modelling of apoptotic signalling allows us to reconstruct a continuous intracellular landscape of sensitivity and tolerance. This provides a basis for investigating reversible tolerance and the recovery of sensitivity at the population level, and their interplay with spatial organisation during tumour spheroid growth. The second investigates plasticity between proliferative and quiescent states in patient-derived rhabdomyosarcoma organoids, where quiescent cells can evade chemotherapy and later re-enter the cell cycle. By combining mechanistic population modelling with mechanistic gene regulatory network inference from time-resolved single-cell transcriptomic data, we investigate the regulation of plasticity, with the aim of identifying potential targets to impair plasticity-mediated resistance. Alongside their biological aims, these projects motivate the longer-term objective of developing continuous phenotype-structured population models grounded in intracellular regulation. I will discuss the methodological challenges involved in pursuing this objective and possible approaches to address them.

• 11h30 – 12h30 : Baptiste Oger

Title: Predicting sugar beet yellows severity and large-scale disease spread from coupled field observations and remote sensing data

Abstract: Beet yellows disease is currently a major threat to sugar beet production in Europe, following recent restrictions on neonicotinoids. Caused by four different viruses, the disease leads to leaf yellowing, reduced photosynthesis, and substantial yield losses. Monitoring its spatial and temporal dynamics remains challenging, as traditional field observations are sparse and costly. In this context, multispectral satellite imagery offers new opportunities to monitor disease severity across multiple spatial scales.
This presentation introduces a two-step approach. First, a statistical framework based on random forests is used to predict beet yellows severity at the field scale from remote sensing data and field observations. Second, a semi-parametric spatio-temporal model is introduced to analyse epidemic dynamics at larger spatial and temporal scales. The proposed approaches are compared with reference methods to assess their predictive performance and ability to capture disease dynamics.
Overall, this work highlights the potential of combining remote sensing and field observations with flexible statistical models for monitoring plant epidemics.

• 12h30 – 14h : Déjeuner
• 14h – 15h : Peter Czuppon

Title: Stochastic models of transposable element dynamics

Abstract:
Transposable elements (TEs) are self-replicating genetic sequences that populate genomes. Their dynamics within a genome are governed by transposition events, where new TE copies are generated, and excision events, which remove TE copies. Moreover, TEs can have deleterious effects on the host, e.g. by inserting into coding genes. This can create negative effects for the host, resulting in TE numbers being counterselected on the population level. In this talk, I will talk about two models to describe TE dynamics. First, we study the mean and variance dynamics arising from individual-based descriptions. We extend the model from [1] to show that TE copy numbers are naturally overdispersed, i.e., the variance TE copy number is larger than the mean in a population and compare these results to data. In a second model, we investigate the interaction of autonomous and nonautonomous TEs, where the latter relies on the former for transposition, similar to a host-parasite system. We describe the stationary copy number distribution of the two TE types in a population.

[1] Pfaffelhuber & Wakolbinger, A diploid population model for copy number variation of genetic elements, Electronic Journal of Probability, 28:1-15, 2023

• 15h – 16h : Oksana Grente

Title: Modelling the effects of lethal control on wolf recolonization dynamics in France

Abstract: The return of wolves to France since the 1990s has led to the resurgence of livestock attacks, particularly on sheep. To prevent livestock losses, the French government has authorized the lethal control of wolves for the past fifteen years. The intensity of this management measure has increased over the past seven years. Here, we aim to model the effects of lethal control on wolf recolonization dynamics in France. Does lethal control accelerate or slow down wolf recolonization? To address this question, we are developing a dynamic multi-state occupancy model using Bayesian inference. We will show the first results of this work in progress. We will discuss the main challenges associated with this modelling approach, including modelling a distribution that is not in an ecological equilibrium and accounting for reciprocal causation between lethal control and wolf presence.