Re: [SOSlib-discuss] Regarding inputting of kinetic rate constants

[Rainer M. <ra...@tb...>]

Hi Roy and Phillip,

You might even get better performance, if you use the integration 
routines at a finer level via the odeModel and integratorInstance 
interface functions.

At this level, you can create both the internal odeModel (derived from 
the SBML input) and the SUNDIALS integrator instance only once, and run 
as many integrations as you like.
The function IntegratorInstance_setVariableValue allows you to set any 
parameters or variables in between (also during) the integration runs.

This level would save you a lot of memory allocation and general 
performance.

You can use Andrew Finney's ParameterScanner.c program in the example 
folder of the SOSlib distribution as a template for your own program, or 
just use it as a commandline tool if you are happy with its output.

It's usage is described at
  http://www.tbi.univie.ac.at/~raim/odeSolver/doc/app.html#param

If you are interested, I can help you changing the output of 
ParameterScanner.c to your requirements. Currently it only gives the 
results of the chosen variable at some specified end-time of 
integration, and that in gnuplot format. Would you need the whole 
time-courses?

Also, if you want to analyze local parameters of kinetic laws, you will 
need to "globalize" these parameters before calling any SOSlib 
functions, because SOSlib only treats global parameters as constants of 
the ODE model, while locals are replaced during ODE construction.
However, we also have such a globalizeParemter function, internally in 
release 1.6.0, but that can easily be "externalized" so you can use it 
in your application (or just copy/paste it :) ).

For the next release we plan much better support for such parameter 
scans also for the main commandline tool, as we realize that this is the 
most widely used application for SOSlib.

Rainer



Roy Wollman wrote:
> Hi Phillip,
> 
> Did you consider using it as a library instead of using the command line 
> utility?
> An example of a library usage that changes a parameter can be found:
> 
> http://www.csc.fi/english/research/sciences/bioscience/programs/sysbio/parameterChange.c
> 
> It will be almost the exact thing to change initial conditions, without 
> parsing the xml file.
> 
> Roy
> 
> 
> Phillip Tabor wrote:
>> Hello,
>>
>> I am attempting to use the odeSolver commandline tool and I'd like to 
>> know if there is a way to input the initial rate constants as an 
>> array. In other words, does the program have to parse the .xml file at 
>> every iteration, or is there a way to override this and simply pass it 
>> an array of numbers to be assigned as rates to the odes?
>>
>> I am attempting to use the software to fit kinetic rate constants with 
>> an evolutionary algorithm. This requires many executions of the 
>> program, and having to parse the .xml file on each run will add a 
>> considerable amount of overhead to the execution time.
>>
>>
>> ------------------------------------------------------------------------
>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys-and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> sbmlsolver-discuss mailing list
>> sbm...@li...
>> https://lists.sourceforge.net/lists/listinfo/sbmlsolver-discuss
>>   
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> sbmlsolver-discuss mailing list
> sbm...@li...
> https://lists.sourceforge.net/lists/listinfo/sbmlsolver-discuss