Dynamical compensation and structural identifiability of biological models: Analysis, implications, and reconciliation

@article{Villaverde2017DynamicalCA,
  title={Dynamical compensation and structural identifiability of biological models: Analysis, implications, and reconciliation},
  author={Alejandro Fern{\'a}ndez Villaverde and Julio R. Banga},
  journal={PLoS Computational Biology},
  year={2017},
  volume={13}
}
The concept of dynamical compensation has been recently introduced to describe the ability of a biological system to keep its output dynamics unchanged in the face of varying parameters. However, the original definition of dynamical compensation amounts to lack of structural identifiability. This is relevant if model parameters need to be estimated, as is often the case in biological modelling. Care should we taken when using an unidentifiable model to extract biological insight: the estimated… 

Figures and Tables from this paper

Observability and Structural Identifiability of Nonlinear Biological Systems
TLDR
This review article is threefold: to serve as a tutorial on observability and structural identifiability of nonlinear systems, using the differential geometry approach for their analysis; to review recent advances in the field; and to identify open problems and suggest new avenues for research in this area.
Full observability and estimation of unknown inputs, states and parameters of nonlinear biological models
TLDR
The results demonstrate the feasibility of an integrated strategy for analysing the theoretical possibility of determining the states, parameters and inputs to a system and solving the practical problem of actually estimating their values.
Repairing dynamic models: a method to obtain identifiable and observable reparameterizations with mechanistic insights
TLDR
This work presents a methodology, AutoRepar, that corrects SIO deficiencies automatically, yielding reparameterized models that are structurally identifiable and observable, and increases the applicability of mechanistic models, enabling them to provide reliable information about their parameters and dynamics.
Input-Dependent Structural Identifiability of Nonlinear Systems
TLDR
A symbolic method is presented that characterizes the input that a model requires to be structurally identifiable and determines which derivatives must be non-zero in order to have a sufficiently exciting input.
The underlying connections between identifiability, active subspaces, and parameter space dimension reduction
TLDR
It is shown that active subspace methods have intrinsic connections to methods from sensitivity analysis and identifiability, and indeed that it is possible to frame each approach in a unified framework.
Identifiability of infection model parameters early in an epidemic
TLDR
It is observed that estimation will fail in practice if only infected case data during the early part of the epidemic (pre-peak) is available, and the existence of the unidentifiability manifold obstructs the ability to exactly determine the parameters.
Symmetries in Dynamic Models of Biological Systems: Mathematical Foundations and Implications
TLDR
An overview of the types of symmetries that appear in dynamic models, the mathematical tools available for their analyses, the ways in which they are related to system properties, and the implications for biological modeling is provided.
Structural identifiability of large systems biology models
TLDR
A computationally efficient algorithm well suited to the analysis of large models, in the model development process is introduced that allows for unknown initial conditions to be parameterised and accordingly, repameterisations requiring the transformation of states associated with unidentifiable initial conditions, can easily be obtained.
Dynamic compensation, parameter identifiability, and equivariances
TLDR
This work shows how dynamical compensation of biological circuits can be formulated in terms of a well-known concept in systems biology, statistics, and control theory—that of parameter structural non-identifiability, and obtains as a special case the sufficient criterion discussed.
...
...

References

SHOWING 1-10 OF 48 REFERENCES
Dynamical compensation in biological systems as a particular case of structural non-identifiability
TLDR
It is argued that the similitude between DC and Structural Identifiability (SI) can be explained in terms of (lack of) the latter, and it is proposed to exploit this fact by using currently existing tools for SI analysis to perform DC analysis.
Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood
TLDR
This work suggests an approach that exploits the profile likelihood that enables to detect structural non-identifiabilities, which manifest in functionally related model parameters, that might arise due to limited amount and quality of experimental data.
Structural Identifiability of Dynamic Systems Biology Models
TLDR
This paper shows how to analyse the structural identifiability of a very general class of nonlinear models by extending methods originally developed for studying observability by implementing STRIKE-GOLDD, a method implemented in a MATLAB toolbox.
Structural Identifiability of Systems Biology Models: A Critical Comparison of Methods
TLDR
The results reveal that the generating series approach, in combination with identifiability tableaus, offers the most advantageous compromise among range of applicability, computational complexity and information provided.
A note on dynamical compensation and its relation to parameter identifiability
TLDR
This work clarifies the apparent discrepancy by requiring that the parameters of DC circuits be identifiable both away from steady-state and when measuring other system variables.
Parameter and structural identifiability concepts and ambiguities: a critical review and analysis.
TLDR
The importance of identifiability analysis prior to performing a parameter estimation experiment, the algebraic difficulties, possible pitfalls, and not-so-apparent ambiguities that may be encountered--even for simple models--and the utility of the concept in experiment design are illustrated.
Comparison of approaches for parameter identifiability analysis of biological systems
TLDR
A case study using three current approaches for structural identifiability analysis for an application from cell biology, discussing strength and weaknesses of each of them and illustrating how they can be applied to real world problems.
Parameter Identifiability of Fundamental Pharmacodynamic Models
TLDR
Issues of parameter identifiability of routinely used pharmacodynamics models are considered and a case study was performed applying an unidentifiable model to real experimental data to show how performing such an analysis prior to parameter estimation can improve the parameter estimation process and model performance.
Delineating parameter unidentifiabilities in complex models.
TLDR
This work describes a scalable method for detecting unidentifiabilities, as well as the functional relations defining them, for generic models, and demonstrates the duality between multiscale sloppiness and the geometry of confidence regions surrounding parameter estimates made where measurement uncertainty is non-negligible.
Higher-order Lie symmetries in identifiability and predictability analysis of dynamic models.
TLDR
A direct search algorithm is established for the determination of admitted Lie group symmetries and it is shown how models with nonidentifiable parameters can still be employed to make reliable predictions.
...
...