Biology Reference
In-Depth Information
One view of systems biology
Experiment
Theory
Computation/
modelling
Technology
Figure 6
Systems biology as an iterative interplay between theory, experiment and
technology development and modelling.
consistency of its behaviour with experimental observations. While these roles
of mathematics may be stronger in systems biology than in other sciences, they
are not qualitatively different.
The special role of mathematics (which we take to include numerical com-
putation) in systems biology derives from the following. It is an aim of systems
biology to understand how properties emerge in the interactions of components
of systems. The emergence of these new properties should be completely deter-
mined by all those interactive properties. If the interaction properties of the
components are correctly known on the basis of experiments with the individual
molecule species, then emergence of the new properties in a precise computer
model is inescapable. The very emergence is thus not in this direct sense sub-
ject to experimental testing. In this aspect systems biology is not subject to
experimental testing either. It may be subject to computational testing, however.
In molecular biology similar situations may arise. The properties of a molecule
are proposed to have an effect on its behaviour, such as that the adjacency of
two glutamate residues in a protein are responsible for the binding of calcium.
Usually in molecular biology no time nor effort is wasted in calculating whether
indeed in principle the adjacency of the two glutamate residues could enhance
calcium binding; this is considered 'obvious' (actually, it may not be quite obvi-
ous; protein dynamics calculations should perhaps be carried out; but in view
of the many nonlinear interactions involved, this is akin to invoking systems
biology). In systems biology it is more often not trivial to see whether a proposed
mechanism for emergence could account for the emerging property, even inde-
pendent of whether the proposed interactive properties are real experimentally. It
involves a computational experiment to check if indeed the proposed interactions
could generate the emergent behaviour. This is so because the interactions are
so complex that an immediate intuitive prediction is impossible, and because the
