In article <33865943.D1BA850 at ix.netcom.com>, rphair at ix.netcom.com says...
>>While it's true that few kinetic data have yet been published for gene
>expression networks, this is about to change dramatically. Large-scale
>gene expression kinetics are already being recorded by a number of
>high-throughput technologies (e.g. ODN chips, and automated RT-PCR).
This is good news. These data are absolutly essential for any model to be
built. They might also help with control analysis...
>In a policy article in Science last year, Eric Lander asserted we should
>be preparing analytical techniques to handle *thousands* of state
>variables. He describes "whole-genome expression studies", and he is not
>talking about the distant future. Several weeks ago, I heard a lecture
>and the US National Academy of Sciences by Lee Hood, who made much the
>same point, although he has in mind what he believes is yet a better
>technology capable, he thinks, of 3 to 5 decades of dynamic range.
I can't wait to see that :) But I also hope someone can start doing thjs in
vivo (we can always ask for more!)
>So my point is, I think the data will be in hand long before we are
>prepared to analyze them. My personal record for number of state
>variables in a published model is less than 50, and even unpublished
>I've never exceeded 100. It seems to me I once had a dream about a
>model with 500 state variables.... or was is a nightmare? :)
Definitly a nightmare! Have you ever tried to fit such a large model (ie the
kinetic constants) to observed data? There is a fundamental problem that
models with so many parameters can fit any data. So we'll be unable to
figure out which is the correct model. That is why we have to be careful
when attacking reductionism... we can't just have reductionist methods but
we sure do need them _as_well_ as holistic ones. So I guess there is a need
for measurements of isolated subcomponents as well...
>Does anyone here have experience with ODE solvers that can handle
>hundreds or even thousands of state variables? Has anyone explored the
>computability questions surrounding such large systems?
This is not a new problem. In physics they have to deal with this all the
time. For example in computational fluid dynamics (CFD) there are usually
very large systems of ODEs to solve. The standard solvers in ODEPACK are
quite capable of handling very large systems. Specifically the one used in
GEPASI and SCAMP, named LSODA, is very good for this. This integrator has an
added advantage in that it measures the stiffness of the equations
throughout the integration and uses either the BDF or Adams methods, if the
equations are stiff or non-stiff respectively (it is a variable step size,
variable order, variable method algorithm).
--
Pedro Mendes Inst. Biological Sciences
prm at aber.ac.uk (home: pedro at enzyme.demon.co.uk) Univ. Wales, Aberystwyth,
http://gepasi.dbs.aber.ac.uk/pedro/prmhome.htm Ceredigion SY23 3DA, U.K.
Author of GEPASI, a biochemical kinetics simulator
http://gepasi.dbs.aber.ac.uk/softw/Gepasi.html
