Biology Reference
In-Depth Information
unexplained but is predicted by a probabilistic theory of embryo-
genesis (Kupiec, 1983). In our computer model, this rate of failure
can be reduced to less than 10% by optimising the values of the
parameters.
If we perform a large number of simulations, we create the
equivalent of a virtual species: a series of structures forms which are
different from each other even though their tissue organisation is
common to them all. This phenomenon is similar to what happens
in a biological species in which all the individuals are different even
though they share the same organisation. In certain species, such as
those of mammals, variability from one individual to another is rel-
atively limited. In other species, such as certain plants, it can be
much greater. Our virtual species is obviously very simple and
scarcely constrained compared with a real species, so its longitudi-
nal variability is great.
The mechanism brought into play here is different from a
reaction-diffusion or self-organisation phenomenon. In these the-
ories, the change of state of a system depends on specific reac-
tions between its components and on differences in their diffusion
speed. Randomness is only involved as a fluctuation setting off
the deterministic dynamics which make the system swing from
one state of equilibrium to another. From the moment the
dynamics are set in motion, there will always be a change in the
state of the system (see Fig. 3). In contrast, the Darwinian model
involves an intrinsically probabilistic mechanism based on non-
specific reactions (cells changing type according to probabilistic
laws) subject to a selective constraint (interdependence for
proliferation). Due to the intrinsically probabilistic character of
this model, from the moment the proliferation of cells is triggered
there is always a certain rate of failure in their forming a cell
bilayer structure. In addition, even though the values of the
parameters of diffusion are identical for the two types of molecule
(which is the case for the examples we are presenting) the sys-
tem is being structured. This structuring does not depend there-
fore on a difference in diffusion, as a reaction-diffusion model
implies.
