Biology Reference
In-Depth Information
4.1. The goals of systems biology
A discussion of what is or should be the methodology of systems biology requires
us to be explicit about our goals in systems biology. The main one, of course, is
to understand more general principles underlying the behaviour and mechanistic
workings of the complete biological systems that sustain life. After all, and as we
discussed above, systems biology should be a science and not just a technology
for analysing special cases. Systems biology should discover new scientific
laws, which may relate as much to physical-chemical, organizational and fitness
aspects as to biochemical principles. With respect to this aim, mathematics
should not take the form of modeling but rather constitute a way of codifying
proposed or verified laws or principles. A case in point is the connectivity
law of metabolic control analysis (see Fell, 1996; Heinrich & Schuster, 1996;
Kell & Westerhoff, 1986; Westerhoff & van Dam, 1987), which can be most
strictly formulated after defining a new property (i.e. the elasticity, see above)
in mathematical terms.
A second aim then is the ability to understand the inner workings of particular
living systems. Ultimately this is best done by having a computational or math-
ematical model of the system in terms of its components and the quantitative
nature of the interactions between them. Such a model could be the result of
'simulation' and 'fitting', the model being adjusted in terms of its structure
and/or its parameter values, until it describes the observed system behaviour.
That description may then constitute understanding. Such a description corre-
sponds to a mechanistic explanation but now in the systems sense.
However, as in other kinds of modelling (Corne et al., 1999; Kell & Knowles,
2006), we want more: A third aim derives from the ability to make predictions
about the possible future behaviour of the system on the basis of changes we
might make to our models. This creates possibilities of further testing the quality
of the model, which is the third aim of modelling. Using a model to make such
predictions forbids its further adjustment whilst calculating the prediction; no
fitting should be involved at such a stage. The same is true in machine learning
(Duda et al., 2001; Hastie et al., 2001; Rowland, 2003). A related, fourth aim
of modelling is the use of the model for technological or therapeutic purposes.
The fifth or ultimate aim of systems biology combines the above; it is the
aim of accomplishing the mission of the life sciences and understanding living
systems in molecular terms, thereby opening such 'applied' avenues as prognosis,
diagnosis, preventive medicine and lifestyle adjustment, therapy, drug design
and biotechnology.
Here we have addressed the understanding of biological systems more than
their explanation in an evolutionary context. Where we addressed explanation
this is in terms of the direct causal mechanisms rather than those that derive
from divergence and selection for fitness or stability or observability. After
all, biological systems live in the absence of evolution. Our discussion has
