Biology Reference
In-Depth Information
1974) as control turned out not to depend on any single step in most metabolic
pathways, but to be distributed. In the larger detailed models of systems biology,
the notion of distributed functionality becomes indispensable.
Pathway modeling and the bottom-up branch of systems biology may be
viewed as constituting a continuous tradition. The change from biological cyber-
netics and systems analysis to top-down systems biology is more dramatic.
Cybernetic modeling proceeds by functional decomposition of a system, whereas
in top-down systems biology this is only one, the less favored, strategy. Decom-
position of a network into modules by dynamical criteria is favored. In addition,
kinetic data of the input-output type, most important for cybernetics, are not so
important for systems biology, where the dynamics of the system is studied by
an analysis of 'omic' data sets. The latter also provide the rich structural data
on which top-down modeling in systems biology relies. We therefore estimate
the link between 'omics' and top-down systems biology, its second root, to be
much stronger than the one between 'omics' and bottom-up systems biology
(Fig. 1). This result is astonishing in a way, because 'omic' data seem to be data
about components rather than systems. But this classification holds within the
framework of 'omics' only. In top-down systems biology, structural 'omic' data
as well as kinetic data obtained from them should be regarded as top-level data
(Palsson, 2006, pp. 63-65).
9
These components are not considered with respect
to their individual role, they merely represent nodes of the network. 'Omic'
data, then, define primarily the state and dynamics of the network as a whole
and are only secondarily analyzed with respect to lower-level modules. A move
away from localized functionality, as was mentioned with respect to bottom-up
systems biology, occurs in the top-down branch as well. Cybernetics assumes
localized functionality, although as a result of the black-boxing of regulatory
pathways the functions are not ascribable to any component if the approach is
not combined with pathway analysis. Depending on the individuation of models
by functional or dynamical criteria, localized functionality is replaced in top-
down systems biology by either distributed functionality or assigning distributed
dynamical roles within the network.
Like the cases from Section 4, and in particular that of the biological clock
model shown, pathway modeling and biological cybernetics are not completely
separable. Modeling here is top-down, but then plausible molecular regulatory
mechanisms are substituting the control elements that are standardly used in
cybernetics. Cybernetic models thus often refer to the molecular basis of regu-
latory circuits and hence incorporate bottom-up models. This means that in fact
both fields are not completely separated but linked to each other, as indicated by
the vertical arrow in the left part of Fig. 1. By this link, pathway modeling also
9
Palsson speaks about top-down data types instead of top-level data, already referring to the direction of
modeling from the systems level down to its (in this case functionally individuated) modules.
