Identifying latent dynamic components in biological systems

Ivan Kondofersky, Christiane Fuchs, Fabian J. Theis

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


In systems biology, a general aim is to derive regulatory models from multivariate readouts, thereby generating predictions for novel experiments. However any model only approximates reality, leaving out details or regulations. These may be completely new entities such as microRNAs or metabolic fluxes which have a substantial contribution to the network structure and can be used to improve the model describing the regulatory system and thus produce meaningful results. In this poster, we consider the case where a given model fails to predict a set of observations with acceptable accuracy. In order to refine the model, we propose an algorithm for inferring additional upstream species that improve the prediction as well as the model fit and at the same time are subject to the model dynamics. In the studied context of ODE-based models, this means systematically extending the network by an additional latent dynamic variable. This variable is modeled by splines in order to easily access derivatives; the influence vector of the variable onto the species is then estimated from the data via model selection.

Original languageEnglish
Title of host publicationComputational Methods in Systems Biology - 11th International Conference, CMSB 2013, Proceedings
Number of pages2
StatePublished - 2013
Event11th International Conference on Computational Methods in Systems Biology, CMSB 2013 - Klosterneuburg, Austria
Duration: 22 Sep 201324 Sep 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8130 LNBI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference11th International Conference on Computational Methods in Systems Biology, CMSB 2013


  • Differential equations
  • Dynamical modeling
  • Maximum likelihood estimation
  • Model selection
  • Splines


Dive into the research topics of 'Identifying latent dynamic components in biological systems'. Together they form a unique fingerprint.

Cite this