Reference : Switch of Sensitivity Dynamics Revealed with DyGloSA Toolbox for Dynamical Global Sen...
Scientific journals : Article
Life sciences : Multidisciplinary, general & others
http://hdl.handle.net/10993/14193
Switch of Sensitivity Dynamics Revealed with DyGloSA Toolbox for Dynamical Global Sensitivity Analysis as an Early Warning for System's Critical Transition.
English
Baumuratova, Tatiana mailto [University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Life Science Research Unit]
Dobre, Simona [> >]
Bastogne, Thierry [> >]
Sauter, Thomas mailto [University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Life Science Research Unit]
2013
PloS one
8
12
e82973
Yes (verified by ORBilu)
International
1932-6203
1932-6203
United States
[en] Systems with bifurcations may experience abrupt irreversible and often unwanted shifts in their performance, called critical transitions. For many systems like climate, economy, ecosystems it is highly desirable to identify indicators serving as early warnings of such regime shifts. Several statistical measures were recently proposed as early warnings of critical transitions including increased variance, autocorrelation and skewness of experimental or model-generated data. The lack of automatized tool for model-based prediction of critical transitions led to designing DyGloSA - a MATLAB toolbox for dynamical global parameter sensitivity analysis (GPSA) of ordinary differential equations models. We suggest that the switch in dynamics of parameter sensitivities revealed by our toolbox is an early warning that a system is approaching a critical transition. We illustrate the efficiency of our toolbox by analyzing several models with bifurcations and predicting the time periods when systems can still avoid going to a critical transition by manipulating certain parameter values, which is not detectable with the existing SA techniques. DyGloSA is based on the SBToolbox2 and contains functions, which compute dynamically the global sensitivity indices of the system by applying four main GPSA methods: eFAST, Sobol's ANOVA, PRCC and WALS. It includes parallelized versions of the functions enabling significant reduction of the computational time (up to 12 times). DyGloSA is freely available as a set of MATLAB scripts at http://bio.uni.lu/systems_biology/software/dyglosa. It requires installation of MATLAB (versions R2008b or later) and the Systems Biology Toolbox2 available at www.sbtoolbox2.org. DyGloSA can be run on Windows and Linux systems, -32 and -64 bits.
http://hdl.handle.net/10993/14193
10.1371/journal.pone.0082973

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Open access
DyGloSA_Baumuratova2013.pdfPublisher postprint1.67 MBView/Open

Bookmark and Share SFX Query

All documents in ORBilu are protected by a user license.