Microbial communities are complex networks of microorganisms that coexist and interact within shared environments. These communities exhibit highly interdependent dynamics that are challenging to study experimentally. Mathematical models, particularly those based on Ordinary Differential Equations (ODEs), are fundamental to modeling these dynamics, providing insights into microbial interactions and facilitating the development of "in silico" experiments that simulate real-world community behavior. However, these models often rely on parameters that are either unknown or difficult to measure experimentally, which introduces uncertainty into the model’s accuracy [1]. To address this, we first conduct a Structural Identifiability Analysis to ensure that the model’s parameters can be uniquely determined from available experimental data. For this purpose, we employ advanced tools such as the GenSSI2.0 Matlab toolbox, and the Julia packages SIAN and StructuralIdentifiability.jl [2]. Ensuring structural identifiability is critical to validating the model’s ability to represent the biological system accurately. Once the model is confirmed to be structurally identifiable, we proceed with a Practical Identifiability Analysis with the Matlab toolbox AMIGO2 [3]. This involves using advanced parameter estimation techniques that combine maximum likelihood estimation with global optimization methods. Overall, this approach bridges the gap between theoretical models and experimental data. The proposed methodology not only advances our understanding of microbial interactions but also offers a pathway for applying these insights to real-world challenges in health, biotechnology, and environmental management

Poster.-- 6th BYMAT Conference: Bringing Young Mathematicians Together, November 4th - 7th, Valladolid, Spain

Peer reviewed