There was a poster at the TCNP meeting from the developers of gfit. It
is a very general package for building models to fit experimental data
via optimization. You basically write a matlab program for your model
and describe a set of experiments (as values for some of the input and
output parameters for your model) and gfit tries to optimize the
unknowns in your model to fid the experimental data. http://gfit.sourceforge.net
It is from the same group as wrote Virtual Cell.
Something we should be aware of.