Uncertainty analysis for non-identifiable dynamical systems: Profile likelihoods, bootstrapping and more

Fabian Fröhlich, Fabian J. Theis, Jan Hasenauer

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

41 Scopus citations

Abstract

Dynamical systems are widely used to describe the behaviour of biological systems.When estimating parameters of dynamical systems, noise and limited availability of measurements can lead to uncertainties. These uncertainties have to be studied to understand the limitations and the predictive power of a model. Several methods for uncertainty analysis are available. In this paper we analysed and compared bootstrapping, profile likelihood, Fisher information matrix, and multi-start based approaches for uncertainty analysis. The analysis was carried out on two models which contain structurally non-identifiable parameters. We showed that bootstrapping, multi-start optimisation, and Fisher information matrix based approaches yield misleading results for parameters which are structurally non-identifiable. We provide a simple and intuitive explanation for this, using geometric arguments.

Original languageEnglish
Title of host publicationComputational Methods in Systems Biology - 12th International Conference, CMSB 2014, Proceedings
EditorsPedro Mendes, Joseph O. Dada, Kieran Smallbone, Pedro Mendes
PublisherSpringer Verlag
Pages61-72
Number of pages12
ISBN (Electronic)9783319129815
DOIs
StatePublished - 2014
Event12th International Conference on Computational Methods in Systems Biology, CMSB 2014 - Manchester, United Kingdom
Duration: 17 Nov 201419 Nov 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8859
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th International Conference on Computational Methods in Systems Biology, CMSB 2014
Country/TerritoryUnited Kingdom
CityManchester
Period17/11/1419/11/14

Keywords

  • Bootstrapping
  • Identifiability
  • Parameter estimation
  • Profile likelihood
  • Uncertainty analysis

Fingerprint

Dive into the research topics of 'Uncertainty analysis for non-identifiable dynamical systems: Profile likelihoods, bootstrapping and more'. Together they form a unique fingerprint.

Cite this