Biology Reference
In-Depth Information
'omic' scope still requires top-down modeling that embeds the modules into the
cellular processes. Such mixed modeling was successful with different metabolic,
genetic-regulatory, and signaling networks (see Kitano, 2002c; Palsson, 2006).
Relying on known functional modules, however, may distort the model of the
network to the extent that modules that are individuated functionally (often on
the basis of intuitive reasoning) may not be as independent from the processes in
which they are embedded as the separate bottom-up models may suggest. Func-
tional modules need not coincide with dynamical modules of a network, hence
an approach that first models functional modules separately and then assembles
these models into a network might end up with a distorted network structure.
To avoid such distortion, a second technique of network decomposition uses
criteria gained from the structure and dynamics of the network itself. Different
mathematical methods are used to individuate so-called unbiased modules, i.e.,
modules as delineated from the embedding network by criteria for strong internal
and weak external interaction rather than by functional considerations (Rohwer
et al., 1996; Koza et al., 2002; Friedman, 2004; Papin et al., 2004). We leave
the analysis of the technical details and pitfalls of 'unbiased modularization' to
the specialists, but will return to some of its consequences in Section 8.1.
This kind of top-down modeling shows certain similarities with biological
cybernetics, although in contrast to the latter it does account for the molecular
components of a network. The strategy of top-down systems biology is to identify
modules within a network, each of which displays a characteristic dynamics.
In biological cybernetics, regulatory instances are postulated within the system,
which are also autonomous to a certain degree, yet interconnected. However,
in systems biology the system is given in terms of data about its components,
while biological cybernetics does not necessarily refer to any component within
the black-boxed system.
7. THE STRUCTURE OF THE FIELD
Having described some of the characteristics of models that can be attributed
to the different roots and branches of systems biology, we now want to use
this material to elaborate the relationships that hold between the different fields.
Although it should be possible in principle to reconstruct a theory net from the
related models and to analyze all the links between them, this would require
reconstructive efforts of almost 'omic' dimensions. We therefore draw a coarser-
grained picture of the links that hold between the different fields, concentrating
on questions about the principal relationships between the various approaches.
The field we are looking at may itself be conceived as a network of the five
approaches that we label pathway modeling, biological cybernetics, 'omics,'
bottom-up systems biology, and top-down systems biology, respectively. The
