Biology Reference
In-Depth Information
emergence depends on the particular magnitude of the parameter values, i.e. on
the particular condition the system is in. (We note, though, that in a sense, such
questions about protein engineering are not quantitative, since changing one or
both of a pair of adjacent glutamates to alanine may perfectly well change the
structure and dynamics of the enzyme irrespective of any effect on their ability
to bind calcium.) It is of course well known that even simple systems can exhibit
very complex dynamics (Abraham & Shaw, 1992; May, 1976). Accordingly,
computation here plays something of the role of experimentation in other sci-
ences. The hypothesis that an experimentally established set of interactions is
responsible for certain emergent behaviour in the system needs to be tested by
performing calculations.
Although this situation is new to much of the life sciences and was not
made very explicit in the original philosophies of physics (Carnap, 1966), it
is standard to present-day physics and chemistry. In particle physics and in
statistical thermodynamics, certain properties may be known experimentally.
The question is then asked whether those properties may be responsible for
certain observed behaviour, and the answer is obtained solely by numerical
experimentation.
We recently carried out this type of numerical experimental systems biology
when proposing that the compound acetaldehyde might be 'responsible' for
the synchronization of glycolytic oscillations between individual yeast cells
(Reijenga et al., 2005a). Putting in the actual structure of the network in so
far as we could, we calculated that the synchronization should indeed occur.
More recently, we posed the hypothesis that the glycolytic oscillations in yeast
are not controlled at a single step such as the proposed pace-maker enzyme
phosphofructokinase, but at many points in the network at the same time. Again
numerical experiments based on what was already known experimentally about
the interaction and networking in the system, served to verify the hypothesis in
the numerical sense (Reijenga et al., 2005b).
We should like to emphasize that in no way do we wish to detract from the
importance of experimental work for systems biology. If anything, experimenta-
tion is more important to systems biology than to molecular biology, in view of
the strong dependence of what actually happens on the precise parameter values.
It is just that mathematics is also more important to systems biology than it is
to molecular biology.
4.4.1. Precision, silicon cells and the calculation of emergence
The calculations we referred to here are often deductive in the sense that they
start from a hypothesis and calculate whether indeed the proposed mechanisms
of emergence deliver the proposed emergent property. However, calculations in
the sense of numerical experiments can also be used to induce general properties.
