Biology Reference
In-Depth Information
be prepared by recombinant means. We shall not dwell here in detail on these
methods, save to note that the systematic production of nominally
all
the pro-
teins of baker's yeast (
S. cerevisiae
) has been performed by Snyder and col-
leagues (e.g. Phizicky et al., 2003; Zhu et al., 2001) and in this sense the
industrialization of such processes has begun (see also, e.g. for
C. elegans
-
http://sgce.cbse.uab.edu/). It is also worth pointing out that even in well-
established recombinant hosts there is a nonlinear interplay between the specifics
of the recombinant vector, the exact host strain and the growth and production
media used to induce the synthesis of the target protein of interest in a form that
allows successful purification and refolding.
The next stage is represented by qualitative binding assay, by which we seek
the 'structural model' that describes the players including substrates, products
and effectors of enzymatic reactions, protein-protein and protein-nucleic inter-
actions and so on (see Fig. 5).
4.4. The special role of mathematics in systems biology: Calculating
emergence
As do most commentators (e.g. Hood, 2003; Ideker et al., 2001; Kitano, 2002;
Naylor, 2004/2005), we (Kell, 2004; Kell, 2005; Kell, 2006; Kell & Knowles,
2006; Westerhoff & Palsson, 2004) consider systems biology to involve an
interplay between theory, computation/modelling and experimental activities.
This interplay is strongly catalysed by the development of new technologies, and
in fact it is these developments more than anything else that has accelerated the
subject (Hood, 2003). It should be noted that Fig. 6 differs rather significantly
from Fig. 3, which we presented as our standard paradigm for scientific activity.
Indeed, we should like to suggest that in systems biology as in other systems
sciences, the role of mathematics is more fundamental than it is in sciences that
deal with single entities of much lower inherent complexity.
Of course, mathematics helps the analyses of the rather complex datasets in
helping to establish correlations, which then feed into the inductive mode of
Fig. 3. It helps ordering the data, then remaining in the empirical box of Fig. 3.
It also helps formulate the hypotheses and theories inside the box theory of
Fig. 3. And it may help deduce experimental implications from the theories,
helping the deductive process depicted in Fig. 3. The reasons for modeling are
numerous, and covered elsewhere (Kell & Knowles, 2006; Klipp et al., 2005),
and include testing whether the model is accurate, in the sense that it reflects,
or can be made to reflect, known experimental facts, analysing the model to
understand which parts of the system contribute most to some desired properties
of interest, hypothesis testing, allowing one to analyse the effects of manipulating
experimental conditions in the model without having to perform complex and
costly experiments, and seeing what changes in the model would improve the
