Athel gave his usual cogent analysis of this question. As one who has
used both of these techniques and has had friendly relations with both
camps, I have a few insights of my own.
1. BSA is essentially an empirical technique. The amplitude coefficient
and the power exponent can be estimated from experimental data and are
valid approximations only over the range of the data. This approach is
useful for estimating uncertainties of a model's predictions.
2. MCA is derived from first principles of biochemical kinetics, and if
the proper rate laws are used should be valid for any concentrations of
the intermediates. Because of its mechanistic basis, this method can
provide molecular level insights into the origins of regulation.
3. Both methods are only valid for small perturbations of the parameters
about the nominal operating point.
4. Eberhard Voit (Medical University of South Carolina, Charleston, SC
USA) has shown that by a change of variables the two methods can be
made mathematicaly equivalent.
5. Which method you choose to use depends on what information is
available and what question you are asking. If you only have
steady-state or time-course concentration data, then you probably will
use BSA. MCA requires detailed kinetic information to identify the true
rate laws and the numerical values of the kinetic parameters. If you
just wish to estimate error envelopes for a simulation, any method of
sensitivity analysis will do. The more realistic the mathematical
representation, the deeper is the possible molecular-level insight.
--
Michael C. Kohn
Laboratory of Computational Biology and Risk Analysis
National Institute of Environmental Health Sciences
P.O. Box 12233, Mail Drop A3-06
Research Triangle Park, NC 27709-2233
Telephone:
919-541-4929 (voice)
919-541-1479 (fax)
e-mail:
kohn at valiant.niehs.nih.gov
Web site:
http://valiant.niehs.nih.gov
