Trois journées par an (Lyon, Marseille et Grenoble) avec 4 exposés.
Journée 2 à Marseille le 21 mars 2025 (10h-16h)
Exposés :
- Claudia Alvarez (Laboratoire de Mathématiques d’Avignon): Stability of the Chemostat system with interaction between species.
The Chemostat system represents a special class of dynamical systems describing the behavior of multiple species that compete over a same substrate within a reactor. This model finds application in diverse areas, notably in the investigation of wastewater treatment. We are interested in the study of a subclass of Chemostat system where direct interspecies interaction is considered. Such interaction can be related to some underlying mutation processes. Our main objective is to understand the stability properties of such systems considering the interaction between species to be as general as possible.
- Florence Bansept (Laboratoire de Chimie Bactérienne, Aix-Marseille): How does intermittent feeding shape the gut microbiome?
Feeding impacts the composition of the gut microbiome in animal hosts. While most research efforts have been directed towards studying the impact of the nutritional bolus on the microbiome composition, the effective use of probiotics suggests that, at least in some cases, the immigration of microbes through feeding also has the potential to modify its composition. In community ecology, immigration is considered a key factor to maintain the diversity of a local community; furthermore, a diverse gut microbiome is usually considered an important determinant of health. Thus, questions arise: do hosts adapt their feeding behavior so as to manipulate microbial immigration in a way that facilitates the maintenance of a diverse flora? What fluctuations are to be expected in the community composition from feeding intermittence, and should sampling be controlled for it?
We develop a mathematical model of microbial communities with birth, death and intermittent migration to study these questions. We show the existence in most cases – when the number of species is large enough – of an optimal feeding strategy, characterized by a relationship between the feeding interval and the food microbial content, that maximizes the average of the microbial diversity over time. We developed a linear approximation of the OFS and showed that its existence is a sufficient condition for the existence of an OFS. In addition, we showed that diversity largely correlates with that of the food, and that the optimal feeding parameters converge to values we are able to explain analytically. We plan to compare our theoretical results with experimental studies in vitro and in different animal hosts, as we expect these effects to depend on host characteristics, like typical transit time.
- Apolline Louvet (BioSP – INRAE): TBA
- Elias Ventre (COMPO – INRIA): Calibration of stochastic models of gene expression from single-cell datasets
Differentiation is the process by which a cell acquires a certain phenotype, through the expression of genes over time. It is now accepted that this dynamic is largely the result of the action of a gene regulatory network (GRN). Studying the action of this GRN in the cells of an organism is one of the main tasks of the field of systems biology. A particularly interesting approach to this problem is to model the dynamics of a cell under the action of its GRN by a stochastic process: GRN inference can then be seen as the calibration of the model from experimental datasets. A first approach is to use a generic model, such as a system of stochastic differential equations, and to employ non-parametric methods based on optimal transport theory. This approach has the advantage of benefiting from highly efficient existing tools. Another approach is to use a mechanistic model, a system of piecewise deterministic Markov processes directly parameterized by a GRN, which is more suitable for taking into account biological priors, but requires the development of appropriate tools. I’ll illustrate the effectiveness of these approaches on simulated and experimental data.
