Forums

Hiroyuki Kuwahara wrote:
> What would be the best way to perform stochastic simulation on a system
> with reactions with non-integer stoichiometries, assuming the system is
> still a Markov process?

It seems to me that there isn't a "good" way of carrying out a discrete
stochastic simulation with non-integer stoichiometries. The derivation
of the standard mass-action stochastic-kinetic rate laws is based on
simple low-order reactions involving integer numbers of molecules. Of
course, if you know the true mechanisms, then you might be able to say
something sensible about the properties of the Markov process, but if
you know that, you can probably re-write the model with integer
stoichiometries...

Now, if the numbers of molecules involved are quite large (but not large
enough to justify a deterministic simulation), then the continuous
diffusion approximation to the discrete-state Markov process might be
acceptable. In this case, the presence of non-integer stoichiometries
and non-standard rate laws isn't such a big deal (at least, no more of a
big deal than it is for deterministic simulations). So simulating a
stochastic diffusion process consistent with the reaction system is
probably OK with non-integer stoichiometries, but is only acceptable in
certain situations.

--
Darren Wilkinson
mailto:darrenjwilkinson@btinternet.com
http://www.darrenjwilkinson.btinternet.co.uk/