Martin Brand wrote
> Can I throw my hat in the ring and compliment Johann Rohwer and
> Jannie Hoffman (Hofmeyr, actually :-) on a clear and useful
> document that formalises the changes in control analysis
> nomenclature that have been building up over the last few years. I
> support the recommendations.
Thank you, Martin.
> In my view, it is the partial response coefficients that really get to
> grips with the things that biochemists want to measure and express when
> they talk about control and regulation. Control coefficients themselves
> are seldom of interest to this group of people, in fact they are confusing
> and off-putting to most biochemists because they answer questions that
> traditional biochemistry (as opposed to biotechnology) doesn't ask and
> isn't interested in. The nomenclature proposed for these partial response
> coefficients is exactly right, and reflects current usage by us and by most
> others.
If you mean what we think you mean then we agree. Biochemists are
traditionally more interested in questions of regulation. Let's
distinguish between external and internal regulators, which, in terms
of the steady state are constant and variable respectively.
If, experimentally, one changes some external regulator and measures
the effect on a steady-state variable then one measures the total
effect. This effect may transmitted via a single step or via a
number of steps (in which case the response measured is the sum of a
number of partial responses). Partial responses are, however, not
directly measurable - one needs, in fact, control and elasticity
coefficients to calculate them.
If the regulator is internal to the system, say an intermediate X
linking a producer and a consumer block (to use your terms), then one
can use top-down or the co-response approach to calculate the block
elasticities. Conceptually, these block elasticities are also
response coeffients of that block seen in isolation, which can be the
sum of a set of partial response coefficients, if X interacts with a
block by more than one route. Again, these partial response
coefficients can only be calculated from elasticities and control
coefficients local to the block.
So, although the partial response coefficients are the ones we want,
we can only measure total response coefficients directly. We can't
escape the control coefficients?
> Although "elasticity coefficient" is horrible we are stuck with it.
> "Sensitivity" was so much better. However, I see no harm in abbreviating
> it to "elasticity" much of the time whilst retaining the full name for
> formal occasions.
Originally, we had "sensitivities" and "elasticities" (control and
elasticity coefficients respectively). We also wish that the clock
can be turned back, but realise that this is impossible. We do
certainly not object to the use of "elasticity" (in fact, we use it
all the time), but for the sake of consistency we feel that formally
we should stick to elasticity coefficients (so that we can speak of
the "coefficients of control analysis" - note that we left out the
"metabolic").
> I would advocate playing down or dropping the word "metabolic" in the title
> "metabolic control analysis". The analysis is much more general than that,
> and the title appears to put off those involved in other related systems
> such as signal transduction or gene expression. I think we want to avoid
> the situation where we coin a new title whenever we apply or extend the
> analysis to these related fields. More generally, control analysis is
> related to the wider field of mathematical analyses of regulation, so we
> need to bear in mind this wider field when we try to harmonise
> nomenclature.
We support dropping the descriptive terms that can precede "control
analysis"; what does everyone else think?
Best regards
Jannie and Johann
Prof. Jannie Hofmeyr Tel: +27 21 808 3039
Dept. of Biochemistry Fax: +27 21 808 3022
University of Stellenbosch e-mail: jhsh at maties.sun.ac.za
Stellenbosch 7600
South Africa
