Physiology-informed regularisation enables training of universal differential equation systems for biological applications
PLOS Computational Biology · 2025
DOI: 10.1371/journal.pcbi.1012198 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS Computational Biology, 2025. Verfasst von Max de Rooij und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands. Hauptthemen laut MEDLINE: Models, Biological, Systems Biology. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Eindhoven AI Systems Institute (EAISI), Dutch Research Council (NWO), Nederlandse Organisatie voor Wetenschappelijk Onderzoek.
Abstract (Originalsprache)
Systems biology tackles the challenge of understanding the high complexity in the internal regulation of homeostasis in the human body through mathematical modelling. These models can aid in the discovery of disease mechanisms and potential drug targets. However, on one hand the development and validation of knowledge-based mechanistic models is time-consuming and does not scale well with increasing features in medical data. On the other hand, data-driven approaches such as machine learning models require large volumes of data to produce generalisable models. The integration of neural networks and mechanistic models, forming universal differential equation (UDE) models, enables the automated learning of unknown model terms with less data than neural networks alone. Nevertheless, estimating parameters for these hybrid models remains difficult with sparse data and limited sampling durations that are common in biological applications. In this work, we propose the use of physiology-informed regularisation, penalising biologically implausible model behavior to guide the UDE towards more physiologically plausible regions of the solution space. In a simulation study we show that physiology-informed regularisation not only results in a more accurate forecasting of model behaviour, but also supports training with less data. We also applied this technique to learn a representation of the rate of glucose appearance in the glucose minimal model using meal response data measured in healthy people. In that case, the inclusion of regularisation reduces variability between UDE-embedded neural networks that were trained from different initial parameter guesses.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands.
- Eindhoven Artificial Intelligence Systems Institute (EAISI), Eindhoven University of Technology, Eindhoven, Netherlands.
- Department of Data Science and Knowledge Discovery, Simula Metropolitan Center for Digital Engineering, Oslo, Norway.
- Amsterdam University Medical Center, University of Amsterdam, Amsterdam, Netherlands.
ORCID-Profile: Max de Rooij, Balázs Erdős, Shauna D O'Donovan
Förderung
- Eindhoven AI Systems Institute (EAISI)
- Dutch Research Council (NWO)
- Nederlandse Organisatie voor Wetenschappelijk Onderzoek
Zitieren
APA
de Rooij, M., Erdős, B., van Riel, N. A. W., & O'Donovan, S. D. (2025). Physiology-informed regularisation enables training of universal differential equation systems for biological applications. PLOS Computational Biology. https://doi.org/10.1371/journal.pcbi.1012198
Vancouver
de Rooij M, Erdős B, van Riel NAW, O'Donovan SD. Physiology-informed regularisation enables training of universal differential equation systems for biological applications. PLOS Computational Biology. 2025. doi:10.1371/journal.pcbi.1012198
BibTeX
@article{deRooij2025,
author = {de Rooij, Max and Erdős, Balázs and van Riel, Natal A W and O'Donovan, Shauna D},
title = {Physiology-informed regularisation enables training of universal differential equation systems for biological applications},
journal = {PLOS Computational Biology},
year = {2025},
doi = {10.1371/journal.pcbi.1012198}
}
EUVIMED