Forums

The issue of parameters might not matter to L2 because most of use can't at
this stage run L2 models in any case. So what ever L2 is decided, those of
use who don't have simulators to run it, would just design to L2, rather
than the other way round as it happenned with L1.

herbert

-----Original Message-----
From: Kouichi Takahashi [mailto:shafi@sfc.keio.ac.jp]
Sent: Saturday, December 07, 2002 12:24 AM
To: Andrew Finney
Cc: 'SBML Discussion List'; Biospice Model Definition Language
Subject: RE: [sbml-discuss] Draft SBML Level 1 version 2 => Level 2



I haven't read the L2 doc very carefully, and I'm a bit confused.
Please point it out if I'm misunderstanding something.

In section 4.6 of the doc:

Parameters defined at the top level are global to the
whole model; parameters that are defined within a reaction are
local to the particular reaction and (within
that reaction) override any global parameters having the same names.

I assume the global parameters are meant for model parameters such as Temp
and pH,
because they can change reaction rates globally, thus can appear in many
rate
equations.

If namespace contamination matters, I think one way is to allow compartment
to
have a parameter list and let them appear in rate equations, though I'm not
sure
this is possible without breaking backward compatibility (if not please just
ignore
this suggestion at least for L2, because probably ProMoT people's modularity
proposal
might did this job elegantly.). Actually Temp and pH
are properties of compartments. E-Cell (especially the version 3), and I
think
ProMoT too, has similar idea although the compartments can be
logical as well as physical.
I think the global model parameters are still useful for modelers.


Herbert:
> > but kinetic
> > parameters are naturally associated with the reaction they
> > belong to and map well with the internal class structures in
> > my software, if the parameters are in the official global
> > list one has to parse the rate laws looking for which rate
> > law each parameter is associated with, not difficult but
> > unnecessary.

For each reaction it is easy; just look up the global parameter list.
>From parameter list, yes, it can be tedious to look for all the rate laws
which are using a particular global parameter, but as a simulator developer
I can't think of need for such operation. Probably you have the need.


regards,
-Kouichi



> IMHO The issue raised by Herbert is perhaps significant for SBML Level 2
> (L2).
>
> Recap: It is suggested that in SBML Level 1 version 2 (L1v2)
> parameters declared in reactions
> are present in the global namespace:
> a parameter declared in a reaction can't have the same
> identifier as a parameter declared anywhere else.
> (this is my interpretation)
>
> [See 3rd paragraph of section 4.8.3 page 22
> and 2nd bullet of section 3.5 page 8 of L2 draft]
>
> I assume that the scope is still restricted:
> a parameter declared in a reaction can't be used
> anywhere else. (otherwise the association
> with the reaction is redundant)
>
> IMHO This is significant for Level 2 because
>
> a) what was a bad idea for L1 is maybe a
> bad idea for L2
>
> b) its not good to have L1v2 as a fork
> in SBML development. I think L2 should
> be a clearly a progression from L1v2.
> This isn't a implementation issue
> as L2 is backwards compatible with
> L1 (mainly because of MathML).
> However it could be confusing
> if L2 was compatible/similar
> to different parts of L1v1 and L1v2.
>
> But do L2 users care???
>
> What do you think?
> Should this change be considered for L2?
>
> yours Andrew
>
> > -----Original Message-----
> > From: owner-sbml-discuss@symbio.jst.go.jp
> > [mailto:owner-sbml-discuss@symbio.jst.go.jp] On Behalf Of
> > Michael Hucka
> > Sent: 06 December 2002 20:33
> > To: SBML Discussion List
> > Subject: Re: [sbml-discuss] Draft SBML Level 1 version 2
> >
> >
> > Herbert sent back an important comment about the
> > interpretation of parameters in SBML L1. With his
> > permission, the message is appended below.
> >
> > We would be very grateful if others could comment on this
> > issue as well. Basically, we have to reach a consensus
> > (which is to say, the most satisfactory and least
> > controversial specification) about the definition of
> > parameters in SBML L1.
> >
> > Mike
> >
> > ------- start of forwarded message (RFC 934 encapsulation) -------
> > From: Herbert Sauro <Herbert_Sauro@kgi.edu>
> > Subject: Level 2
> > Date: Fri, 6 Dec 2002 09:41:30 -0800
> >
> > Mike, you might be aware of this but Gepasi, my software and
> > netbuilder and probably others (eg DBSolve and Luigi's
> > simulator) use a slightly different interpretation for the
> > parameters associated with rate laws than is given in the
> > Level 1 docs. In the official docs it says these are local
> > but the way the rest of us them is in a global sense, it just
> > didn't turn out to be convenient to use them as indicated in
> > the docs. The reasons why I do it this way (don't know why
> > the others do it, you'll have to ask them), but kinetic
> > parameters are naturally associated with the reaction they
> > belong to and map well with the internal class structures in
> > my software, if the parameters are in the official global
> > list one has to parse the rate laws looking for which rate
> > law each parameter is associated with, not difficult but
> > unnecessary. I imagine that since Level 1 will go the way of
> > the dinosaur with Level 2 almost there, it shouldn't be too
> > much of an issue, I won't be using level 2 for a long time
> > since I don't have a simulator that is compatible with it.
> >
> > Herbert
> >
> > - ----------------------------------------------------------
> > There are only three kinds of people in this world; those
> > that can count, and those that can't
> >
> > Dr Herbert M Sauro
> > Keck Graduate Institute
> > 535 Watson Drive
> > Claremont
> > CA, 91711
> > Tel: USA 1-909-607-0377
> > Email : hsauro@kgi.edu <mailto:hsauro@kgi.edu>
> > Web: www.sys-bio.org <http://www.sys-bio.org>
> >
> > ------- end -------
> >
>
>
>