Forums

Dear SBML Community

We often come across situations where it would be advantageous to have more
than one set of parameters included in a SBML model. This makes it easy to
store snapshots of parameter regimes in which a model shows different
behavior.

In one of the SBML meetings in Tokyo we discussed this many years ago,
however there did not seem to be a consensus on how to store this sort of
information. What people ended up doing was to store the information in
proprietary annotations. As I was facing this problem again recently, I
ended up writing down the format of the annotation here:

http://precedings.nature.com/documents/4233/version/1

Perhaps, if there is enough interest we could think about building a Level 3
package round it, or perhaps the groups package will be the right place for
this sort of information.

The reason for me posting this information here, is of course that I'm
interested in seeing support for an annotation like this in other software
tools.

Cheers
Frank

____________________________________________________________
To manage your sbml-discuss list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss

For a web interface to the sbml-discuss mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the sbml-discuss list,
contact sbml-team@caltech.edu

JDesigner stores multiple parameter sets relative to a single model (all
in one file). SBML files are verbose as it is so I didn't want to
repeatedly store the same model multiple times. I know there are users
who use multiple parameter sets in JDesigner quite extensively. Each set
has an Id, the parameters (including initial conditions) with values and
a notes section. Parameter sets can be restored or saved as need be.

Herbert

Lucian Smith wrote:
> * Frank Bergmann <fbergman@u.washington.edu> [2010-02-24 21:16] writes:
>
>> Dear SBML Community
>>
>> We often come across situations where it would be advantageous to have more
>> than one set of parameters included in a SBML model. This makes it easy to
>> store snapshots of parameter regimes in which a model shows different
>> behavior.
>>
>> In one of the SBML meetings in Tokyo we discussed this many years ago,
>> however there did not seem to be a consensus on how to store this sort of
>> information. What people ended up doing was to store the information in
>> proprietary annotations. As I was facing this problem again recently, I
>> ended up writing down the format of the annotation here:
>>
>> http://precedings.nature.com/documents/4233/version/1
>>
>> Perhaps, if there is enough interest we could think about building a Level 3
>> package round it, or perhaps the groups package will be the right place for
>> this sort of information.
>>
>
> The obvious candidate to store this information for me is the Hierarchical
> Modeling package. It won't store everything in one file, but it would
> make it possible to store a series of 'wrapped' models, each with a
> different parameter set, but based on the same underlying model.
>
> Would that be sufficient for your envisioned use cases, or is there a
> distinct advantage to putting multiple versions of a model in a single
> file?
>
> -Lucian
> ____________________________________________________________
> To manage your sbml-discuss list subscription, visit
> https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss
>
> For a web interface to the sbml-discuss mailing list, visit
> http://sbml.org/Forums/
>
> For questions or feedback about the sbml-discuss list,
> contact sbml-team@caltech.edu
>


--
---
Herbert M. Sauro, PhD
Associate Professor
Department of Bioengineering

University of Washington
William H. Foege Building
Box 355061
Seattle, WA 98195-5061
206.685.2119 (tel)
206.685.3300 (shared fax)
http://www.sys-bio.org

____________________________________________________________
To manage your sbml-discuss list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss

For a web interface to the sbml-discuss mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the sbml-discuss list,
contact sbml-team@caltech.edu

Hello Lucian,

On Wed, 24 Feb 2010 21:41:05 +0000
Lucian Smith <lpsmith@spod-central.org> wrote:

> The obvious candidate to store this information for me is the
> Hierarchical Modeling package. It won't store everything in one
> file, but it would make it possible to store a series of 'wrapped'
> models, each with a different parameter set, but based on the same
> underlying model.
>

This is not true. In the hierarchical modeling package the model can be
internal or external. Thus everything in a single file is possible.

Thanks,
Stefan


--
Stefan Hoops, Ph.D.
Senior Project Associate
Virginia Bioinformatics Institute - 0477
Virginia Tech
Bioinformatics Facility II
Blacksburg, Va 24061, USA

Phone: (540) 231-1799
Fax: (540) 231-2606
Email: shoops@vbi.vt.edu
____________________________________________________________
To manage your sbml-discuss list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss

For a web interface to the sbml-discuss mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the sbml-discuss list,
contact sbml-team@caltech.edu

Hello Nicolas,

As said before ... the multiple parameter sets annotation fulfills a need
right now ... in level 2 ... I'm perfectly aware of alternatives in Level 3.


As you know I'm especially interested in SED-ML. And this is in fact one
application for the multiple parameter set annotations. By simply marking
different parameter sets in applications like JDesigner (my library supports
these annotations and transforms them into the multiple parameter sets
annotation), or several other applications, these multiple parameter sets
will be transformed into SED-ML model and ChangedModel constructs, and
simplify the procedure of creating simulation experiments that showcase the
behavior of the model under the different sets.

Cheers
Frank

> -----Original Message-----
> From: sbml-discuss-bounces@caltech.edu [mailto:sbml-discuss-
> bounces@caltech.edu] On Behalf Of Nicolas Le novère
> Sent: Wednesday, February 24, 2010 1:56 PM
> To: SBML Discussion List
> Subject: Re: [sbml-discuss] Managing Multiple Parameter Sets in SBML
>
> Hi,
>
> I strongly believe that all initial conditions and parameter values will
> eventually have to be stripped of SBML. But that won't happen before
> Level4. Meanwhile, several mechanisms have been proposed so far. Lucian
> pointed out one possible way, i.e. the model composition package. Another
> one is the distributions and sets package. We will have to implement this
> package soon anyway, if we want to reach the drug discovery community
> (where the initial conditions are distributions). A third way is to use
SED-
> ML, i.e. to have a reference model, and to alter the values before running
> the simulation steps.
>
> Frank T. Bergmann wrote:
> > Dear SBML Community
> >
> > We often come across situations where it would be advantageous to have
> > more than one set of parameters included in a SBML model. This makes
> > it easy to store snapshots of parameter regimes in which a model shows
> different
> > behavior.
> >
> > In one of the SBML meetings in Tokyo we discussed this many years ago,
> > however there did not seem to be a consensus on how to store this sort
> > of information. What people ended up doing was to store the
> > information in proprietary annotations. As I was facing this problem
> > again recently, I ended up writing down the format of the annotation
> here:
> >
> > http://precedings.nature.com/documents/4233/version/1
> >
> > Perhaps, if there is enough interest we could think about building a
> > Level 3 package round it, or perhaps the groups package will be the
> > right place for this sort of information.
> >
> > The reason for me posting this information here, is of course that I'm
> > interested in seeing support for an annotation like this in other
> > software tools.
> >
> > Cheers
> > Frank
> >
> > ____________________________________________________________
> > To manage your sbml-discuss list subscription, visit
> > https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss
> >
> > For a web interface to the sbml-discuss mailing list, visit
> > http://sbml.org/Forums/
> >
> > For questions or feedback about the sbml-discuss list, contact
> > sbml-team@caltech.edu
>
>
> --
> Nicolas LE NOVERE, Computational Neurobiology, EMBL-EBI, Wellcome-
> Trust Genome Campus, Hinxton CB101SD UK, Mob:+447833147074,
> Tel:+441223494521
> Fax:468,Skype:n.lenovere,AIM:nlenovere,MSN:nlenovere@hotmail.com(N
> OT email) http://www.ebi.ac.uk/~lenov/,
> http://www.ebi.ac.uk/compneur/, @lenovere
>
>
> ____________________________________________________________
> To manage your sbml-discuss list subscription, visit
> https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss
>
> For a web interface to the sbml-discuss mailing list, visit
> http://sbml.org/Forums/
>
> For questions or feedback about the sbml-discuss list,
> contact sbml-team@caltech.edu

____________________________________________________________
To manage your sbml-discuss list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss

For a web interface to the sbml-discuss mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the sbml-discuss list,
contact sbml-team@caltech.edu

* shoops <shoops@vbi.vt.edu> [2010-02-24 22:05] writes:
> Hello Lucian,
>
> On Wed, 24 Feb 2010 21:41:05 +0000
> Lucian Smith <lpsmith@spod-central.org> wrote:
>
> > The obvious candidate to store this information for me is the
> > Hierarchical Modeling package. It won't store everything in one
> > file, but it would make it possible to store a series of 'wrapped'
> > models, each with a different parameter set, but based on the same
> > underlying model.
> >
>
> This is not true. In the hierarchical modeling package the model can be
> internal or external. Thus everything in a single file is possible.

We're starting to drift off-topic from Frank's original question, but what
I meant was that since the Hierarchical Modeling package for L3 is meant
to be 'flattenable' to a single L2/L3 model, you could not encode multiple
initial parameter values in a single SBML file. (At least, as I
understand it?) One might implement it as:

File 1:
The full model, with the first set of initial conditions.
File 2:
A reference to File 1.
Replacements for any element you wanted new initial conditions for
File 3:
A reference to File 1.
Replacements for any element you wanted new initial conditions for

Etc. Each file apart from the first would be pretty small, but you'd
still need multiple files.

I think if you wanted multiple models in a single file (and wanted to use
HM to do it), you'd have to hack the system--perhaps a base-level model,
a model above that that replaced some bits, a model above *that* that
replaced the same bits, etc. If you flattened it, you'd only get the
'top-level' model, but you could abuse the system by pulling out the inner
ones for this purpose. I think that's probably the wrong thing to do, but
would be willing to be convinced otherwise.

-Lucian
____________________________________________________________
To manage your sbml-discuss list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss

For a web interface to the sbml-discuss mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the sbml-discuss list,
contact sbml-team@caltech.edu

Hello Dagmar,

You are correct, the annotation does not address it. As far as the current
implementation is concerned with one click of a save button all parameters
of a model are stored as a new parameter set. This is mainly due to the
fact, that I would not like the user to have to choose which of his 300
favorite parameters he wants to consider to save.

As far as the implementation is concerned a parameter set could just as well
be a subset of parameters, or even just one parameter that could be applied
to the model. A parameter containing only of a subset of parameters is
considered equal to the set of all initial values defined for the model with
only the parameters in the subset being changed.

Cheers
Frank

> -----Original Message-----
> From: sbml-discuss-bounces@caltech.edu [mailto:sbml-discuss-
> bounces@caltech.edu] On Behalf Of Dagmar Waltemath
> Sent: Thursday, February 25, 2010 12:14 AM
> To: SBML Discussion List
> Subject: Re: [sbml-discuss] Managing Multiple Parameter Sets in SBML
>
> Frank, I might have overread it in your proposal, but is your parameterSet
a
> complete set of all parameters?
> Or is the policy that a set can contain a subset of all model's
parameters, and
> only those are changed if the set is active?
>
> Dagmar
>
>
> Frank T. Bergmann wrote:
> > Dear SBML Community
> >
> > We often come across situations where it would be advantageous to have
> > more than one set of parameters included in a SBML model. This makes
> > it easy to store snapshots of parameter regimes in which a model shows
> different
> > behavior.
> >
> > In one of the SBML meetings in Tokyo we discussed this many years ago,
> > however there did not seem to be a consensus on how to store this sort
> > of information. What people ended up doing was to store the
> > information in proprietary annotations. As I was facing this problem
> > again recently, I ended up writing down the format of the annotation
here:
> >
> > http://precedings.nature.com/documents/4233/version/1
> >
> > Perhaps, if there is enough interest we could think about building a
> > Level 3 package round it, or perhaps the groups package will be the
> > right place for this sort of information.
> >
> > The reason for me posting this information here, is of course that I'm
> > interested in seeing support for an annotation like this in other
> > software tools.
> >
> > Cheers
> > Frank
> >
> >
> __________________________________________________________
> __
> > To manage your sbml-discuss list subscription, visit
> > https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss
> >
> > For a web interface to the sbml-discuss mailing list, visit
> > http://sbml.org/Forums/
> >
> > For questions or feedback about the sbml-discuss list, contact
> > sbml-team@caltech.edu
> >
>
> __________________________________________________________
> __
> To manage your sbml-discuss list subscription, visit
> https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss
>
> For a web interface to the sbml-discuss mailing list, visit
> http://sbml.org/Forums/
>
> For questions or feedback about the sbml-discuss list, contact sbml-
> team@caltech.edu

____________________________________________________________
To manage your sbml-discuss list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss

For a web interface to the sbml-discuss mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the sbml-discuss list,
contact sbml-team@caltech.edu