SimBiology

[Jerry]

Hello,

There is a model that have 20 species and 25 parameters. When I ran the model and monitored the dynamic of a species (let's call it "S"), the shape of the signal is great but there are two problems that should be fixed. First, the "S" shows a transient signal between 0 and 20 hours, but the dynamic of "S" should be between 0 and 2 hours. Second, the concentration of "S" shows 10,000 time more than it should be. I need either to adjust one/few parameter(s), or to estimate a missing parameter to fix these problems. Please let me know whether or not there is a specific function within SimBiology allowing me to fix these problems.

Thanks,
Jerry

[Arthur Goldsipe]

Hi Jerry,

If I understand your problem correctly, you are trying to figure out how to change your SimBiology model so that the simulation results better match your expectations. I can think of two features in SimBiology to help in this situation: sensitivity analysis and parameter estimation. The following example illustrates both of these:

http://www.mathworks.com/products/simbiology/examples.html?file=/products/demos/shipping/simbio/gprotein.html

To go in a little more detail on the use of sensitivity analysis, it sounds like you are most interested in how things affect species S. So you could run a sensitivity analysis simulation with S as your sensitivity output and all the parameters (and probably species initial concentrations) as your sensitivity inputs. If you see a sensitivity with a large magnitude over the entire simulation, that might identify a good parameter or initial concentration to adjust in order to decrease the magnitude of S. If you see a sensitivity that changes after 2 hours, that might identify a good parameter or initial condition to adjust in order to cause the dynamics of S to occur at later times.

After you've identified the parameters and initial conditions of interest, you can either manually try varying their values or you can try to estimate the parameters using the function sbioparamestim. This function requires you to have a time course that you are trying to fit the data to, so if you don't have real experimental data you're trying to fit, you'll need to come up with synthetic data that represents the dynamics you are trying to reproduce. Then, you can try to estimate the parameters and initial conditions that you identified as important from sensitivity analysis.

Don't be surprised if the parameter estimation doesn't immediately converge to a good set of of parameters. This is a very difficult optimization problem that is often very sensitive to your initial guesses for the optimal parameter values. If you need to explore a large range of parameter values, you may find it useful to use one of the optimization methods provided to sbioparamestim via the Global Optimization Toolbox (pattern search, genetic algorithm, and particle swarm optimization).

-Arthur

"Jerry " <jerry...@gmail.com> wrote in message <kmg5j2$9i0$1...@newscl01ah.mathworks.com>...

[Jerry]

Hello,

In SimBiology, I need to show that [A] + [A]--> [2A], however, it seems that I couldn't have two species with the same name.

Thanks,
Jerry

[Ingrid Tigges]

Dear Jerry,

Which MATLAB release are you using? In MATLAB R2013a for example the following works as reaction
A + A -> 2 A
If 2A should be a new species the reaction would be
A + A -> [2A]
If you have any further questions regarding this, please let me know.

[Jerry]

Thank you for your suggestion! I used Matlab 2012b. But based on your recommendation, I just installed Matlab 2013a. Once I created species "A" and I wanted to create another "A", I got this error: "The specified name is being used by another species. Specify a different name."

Thanks,
Jerry

[Ingrid Tigges]

Hi Jerry,

I am not sure whether I understand your question precisely.
All references to "A" refer to the same species. Did you try the line of code I mentioned? It does not matter whether you are using MATLAB R2012b or R2013a for this case.

Ingrid

[Jerry]

Hello,

Actually, I created a new sbproj file, and I was trying to add the species ("A" and again "A") plus reactions in the model. I didn't try writing of codes in m.file.

Thanks,
Jerry

[Arthur Goldsipe]

Hi Jerry,

You can create a reaction "A + A -> [2A]" from either the command-line or from the SimBiology Desktop. In both cases, you only need to create a single species named A.

You cannot never have two species named A in the same compartment of a SimBiology model. But that's ok, because in this reaction, both A's really refer to the same species.

Perhaps it would be clearer if you wrote the reaction as "2 A -> [2A]". This should produce equivalent results.

-Arthur