Project properties
Title Predicting Toxicity dynamics in bioprocess design and optimization
Group Systems and Synthetic Biology
Project type thesis
Credits 36
Supervisor(s) Ivan Martin Martin, Jasper Koehorst, Maria Suarez Diez
Examiner(s) Prof. Maria Suarez Diez (SSB) Dr. Rob Smith
Contact info Robert1.smith@wur.nl
Begin date 2025/12/01
End date
Description Microbial cell factories (MCF) are important to design biobased synthesis processes to substitute fossil fuel based production procedures. However, the implementation and scale-up of these bioprocesses is technically challenging, since bioreactor operating conditions largely affect the production of microbial cell factories.
Certain microbial dynamics are particularly complex to model in the context of bioprocesses, e.g. due to the requirement of kinetic constants. An example is the production of a microbial compound that becomes toxic above certain concentrations.
Mechanistic approaches (such as dynamic Flux Balance Analysis, dFBA) can be employed to model these toxicity dynamics when the kinetic constants are available [1] – which is not always the case. Instead, large-scale statistics (i.e. Machine Learning) can also be employed to model and predict the production of a (potentially) toxic compound, when experimental data are available [2].
This project aims to design a neural ODE framework to model toxicity dynamics for a given production process by a MCF in a bioreactor. Results will be compared with the modelling performance of a mechanistic approach based on dFBA (already available) and with other experimental results (to be found) in literature.
The thesis will most likely involve to (1) find a relevant dataset as experimental benchmark, (2) to code (in Python) the indicated neural ODE model, (3) to generate results for a given MCF in a set of scenarios, (4) to compare these results with the experimental datasets and the indicated mechanistic approach, (5) to draw some conclusions on why it works (or not) and what to improve.

References related to project: [1], [2]
Used skills
Requirements Programming in Python. Basic knowledge of Ordinary Differential Equation (ODE) systems, Machine and Deep Learning (i.e. courses SSB30306 or FTE35306). Knowledge of Unix/Linux operating system is also welcomed but not required.