2024-07-09 –, TU-Eindhoven 0.242
Computational models are a valuable tool to study dynamic interactions and the evolution of systems behavior. Our hands-on and interactive workshop will demonstrate how personalized models can be more rapidly generated in Julia using various SciML packages combined with custom implementations. We will cover the implementation of ODE models in Julia, parameter estimation and model selection strategies including parameter sensitivity and identifiability analysis.
Computational models offer a valuable tool for understanding the dynamic interactions between different biological entities, especially in biomedical applications. Personalizing these models with data can shed light on interindividual variation and project future health risks. However, model generation can be computationally expensive. Our hands-on and interactive workshop will demonstrate how personalized models can be more rapidly generated in Julia. We will be mainly using DifferentialEquations.jl combined with Optimization.jl and custom implementations of sensitivity and identifiability analysis approaches. Useing an in-house model of the glucose-insulin system, we will cover the implementation and resolving of ODE systems in Julia, including importing in SBML. We will provide a guide on model selection including parameter sensitivity and identifiability analysis, highlighting efficiencies that can be achieved using Julia. Additionally, we will discuss strategies for parameter estimation, including the benefits of regularization, using a publicly available data set of meal responses. Short presentation will be used to provide necessary background and theory and all methods will be implemented in a Jupyter notebook to facilitate independent learning.
I’m a PhD candidate in systems biology for metabolic disease at the Department of Biomedical Engineering at Eindhoven University of Technology. I am working on model personalisation with scientific machine learning.
Assistant Professor in Modelling and AI for precision nutrition; Department of Biomedical Engineering; Eindhoven University of Technology.
Natal van Riel is Professor of Biomedical Systems Biology at the department of Biomedical Engineering at Eindhoven University of Technology, where he leads the Computational Biology group and the Systems Biology and Metabolic Diseases research program. He is also part-time Professor of Computational Modelling at Amsterdam University Medical Centers (location AMC, University of Amsterdam's Faculty of Medicine). His research focuses on modelling of metabolic networks and physiology, machine learning for parameter estimation, methods for analysis of dynamic models, and applications in Metabolic Syndrome and associated diseases such as Type 2 Diabetes.