This repository contains the code and analysis accompanying my master's thesis:
"Data-Driven Modeling of Metabolic Oscillations in Yeast Cells"
Department of Biochemistry and Molecular Biology
University of Southern Denmark
Author: Henrik Helge Gundestrup
Supervisor: Daniel Wüstner
The thesis investigates metabolic oscillations in yeast using time-lapse fluorescence imaging data. Several modeling approaches are compared for their ability to reconstruct and predict dynamics, including:
- Physics-Informed Neural Networks (PINNs)
- Sparse Identification of Nonlinear Dynamics (SINDy)
- Dynamic Mode Decomposition (DMD)
- Long Short-Term Memory Networks (LSTM)
The models are trained and evaluated on experimental data from Wüstner et al. (SDU), specifically intensity traces related to NAD(P)H metabolism under varying glucose conditions.
Experimental data used in this project will be made available upon reasonable request.
Please contact: [email protected]