This package provides tools for simulation and statistical analysis of systems modelled by kinetic equations. Reaction kinetics can be expressed as systems of ordinary differential equations (ODEs), thus allowing Bayesian estimation of the involved system parameters, such as reaction rates and transition states, and Bayesian selection among candidate models of the unerlying chemical processes or pathways.

At the moment, the package implements the simulation of time courses of the kinetic equations, as well as the simulation of data of the states. Bayesian inference and model selection will be implemented as soon as the prerequired population MCMC algorithms are set up in a parallel fashion.

As a short tutorial, the states and their initial conditions, the model
parameters and their values needed for simulation, and the system of ODEs
defining the kinetic equations are all provided in a file. An example is
available in `test/simulateFitzhughNagumoODEs.jl`

, which demonstrates how to
define the Fitzhugh Nagumo differential equations. The file's sections are
marked as `*** MODEL STATES ***`

, `*** MODEL PARAMETERS ***`

and
`*** MODEL ODES ***`

. The first three stars indicate that the line represents a
title and the sections are distinguished by using one of the self-explantory
strings `MODEL STATES`

, `MODEL PARAMETERS`

or `MODEL ODES`

. Any other trailing
characters are optional.

To parse the file, use

```
using ChemicalKinetics
odeModel = OdeModel("fitzhughNagumoModel.txt")
```

and to set up the tailored CVODE Sundials solver in order to simulate a time course of 25 minutes run

```
ckCvode = CkCvode(odeModel, [i for i in 0:1.:25])
```

Then simulate the ODE system and generate data from it with noise variance equal to 0.1:

```
ode_simulation, data_simulation = ckCvode.simulate_data(0.1)
```

The following plot shows an example of such simulation:

https://github.com/scidom/ChemicalKinetics.jl/blob/master/test/fitzhugh_nagumo_simulation.png

06/20/2013

3 months ago

17 commits