/parameter_estimation_reaction_rates

Matlab implementation of automated parameter estimation method for chemical reaction networks from observed time-series experimental data of reaction rates

Primary LanguageMATLABMIT LicenseMIT

This repository contains the Matlab library corresponding to the manuscript "Parameter estimation for models of chemical reaction networks from experimental data of reaction rates" published in "International journal of control". The manuscript is available at "https://www.tandfonline.com/doi/full/10.1080/00207179.2021.1998636".

The MATLAB library uses the information of a given CRN and the available observed time-series experimental data of reaction rates provided by the user to automatically determine the best-fitting values of the parameters contained in the corresponding mathematical model. 

We have provided two real-life computaional models of biologial processes retrieved from the BioModels database (see, [1] and https://www.ebi.ac.uk/biomodels/). For each of these models we first generate data corresponding to the overall outgoing reaction rates with initially provided values of parameters involved in the model. Next, we perturbe these data with a white Gaussian noise with zero mean and sfficiently small standard deviation. We then apply our parameter estimation technique to determine the best-fitting values of parameters in a fully automated model.

The first one of these realistic examples is a model of Ryanodine Receptor Adaptation (RyRA) (see, [2]), which is a model of a network governed by mass action kinetic rate law. The second example is a model of Protein Kinease Cascades (PKC) (see, [3]), which is a model of a network governed by reversible Michaelis-Menten kinetics. The schematic representations of these two models can be found in the file named "Schematic" located in the folder "Figures". Since the governing laws of these two examples are different, we consider the parameter estimation problem for each of these examples separately. 

The Matlab library uses the information of a given biochemical reaction network provided by the user to automatically estimate the parameters involved in the corresponding mathematical model.

In order to proceed with the parameter estimation, the user needs to:
        
    1. For the model of RyRA.
     
        a. Open the files located in the folder "scr/RyRA" 
        b. Run the file "GH_RyRA.m"
        c. Run the file "GH_Available_Experimental_Data.m"
        d. Run the file "GH_Parameter_Estimation_Procedure.m"
        
    2. For the model of PKC.
     
        a. Open the files located in the folder "scr/PKC" 
        b. Run the file "GH_PKC.m"
        c. Run the file "GH_Available_Experimental_Data.m"
        d. Run the file "GH_Parameter_Estimation_Procedure.m"
              
The automatic parameter estimation procedure provides the following outputs:

     1. The generated time-series data of overall outgoing reaction rates
     2. The mathematical model with the best-fitting values of parameters 
     3. The comparison of the time-series data of overal outgoing reaction rates and the corresponding model-predicted valeus  
     
REFERENCES

[1]  Nicolas Le Novere et al. 
     Biomodels Database: a Free, Centralized Database of Curated, Pub647 lished, Quantitative Kinetic Models of Biochemical and Cellular Systems. 
     Nucleic Acids Research, 34:D689–D691, 2006.

[2]  Joel Keizer and Leslie Levine. 
     Ryanodine Receptor Adaptation and Ca2+ (-) Induced Ca2+ Release-dependent Ca2+ Oscillations. 
     Biophysical Journal, 71(6):3477–3487, 1996.
     
[3]   Nick Markevich, Jan Hoek, and Boris Kholodenko. 
      Signaling Switches and Bistability Arising from Multisite Phosphorylation in Protein Kinase Cascades. 
      The Journal of Cell Biology, 164(3):353–359, 2004