Microbial communities are fundamental for human wellbeing, robust interactions generate a barrier against pathogens and equilibrated microbiota are crucial for human immune balance. In contrast, microbiome dysbiosis is associated with inflammation and symptom aggravation. Emergent properties of the interaction network are likely determinant drivers for human health and microbiome equilibrium.
This project will address three central research questions: First, we ask how community organizations translate into healthy microbiome function, and if patterns in the microbial interaction networks can discriminate healthy from unhealthy microbiomes. Second, how to implement an ecologically sound, mathematical framework that enables to infer community organization from data, in order to be able to overcome the gap between simplistic mechanistic models and complex large-scale microbiome data. Third, how to build on the resulting knowledge of organization patterns for devising targeted intervention strategies (bacteriotherapy).
We will combine complementary expertise to tightly integrate the development of mathematical tools, data analysis and experimental testing for advancing theory on complex microbiome dynamics. In a pilot, we demonstrated that clustering of abundance profiles together with the resolution of microbiome interaction structure are key for designing functional bacterial consortia. Building from there, we will develop a suitable network inference framework using semi-parametric methods and dedicated dimension reduction to accommodate model complexity and data noise limiting prevailing approaches.
With these, we will study empirical data in acute inflammation (lung) and in inflammation-free colonization (gut) and use network science, time series analysis and machine learning to define shifts in interaction patterns that drive pathologic microbial activity. To examine our theoretical insights, we aim to design and implement a bacterial consortium that counteracts gut pathogen invasion by modeling underlying interaction processes.
We will use a mouse model for gut microbiota barrier effect against Vancomycin Resistant Entrococcus (VRE) invasion to conduct perturbation and rescue experiments. Data will be generated to feed the theoretical work and for validating predictions about interactions and metabolic activities. An optimized microbial consortium with anti-VRE effect will be proposed, as well as complementary experiments to unravel the molecular mechanisms.
We consider this model-driven, interdisciplinary research a milestone for achieving targeted microbiome management that will maximize health benefits, reduce the world-wide consumption of antibiotics and provide theory for innovative biomedical strategies.
The research partners are Stefanie Widder (PI, MedUni, Vienna) and David Berry (Uni Vienna) in Austria, Beatrice Laroche (PI, MaIAGE, INRAE) and Pascale Serror (CPE, INRAE) in France.