Biomedical Engineering Reference
In-Depth Information
limits to the evolution of modularity, and neural network models have been used
to explore how modularity can lead to more efficient task management (6). For a
comprehensive review of the recent literature in the area of modularity see (53).
3.4. Purging—Antiredundancy
Whereas redundancy buffers the effect of perturbation, purging acts in the
opposite fashion—amplifying the effects of perturbation—so as to ensure the
purity of a population (29,31). A gene
x
on a genetic background
y
+
z
is func-
tionally antiredundant when the target gene is one of at least two or more genes
(
x
+
y
) contributing to the phenotype epistatically, and, when removal of gene
x
leads to a greater perturbation in the presence of
y
than in the absence of
y
:
f
(
x
,
y
+
z
)
f
(
z
) >>
f
(
y
+
z
).
Purging is only effective when individual replication rates are sufficiently
large to tolerate the effects of removal of defective components. Thus apop-
tosis—programmed cell death—is a common strategy for eliminating cells upon
damage to their genomes or upon infection, provided these cell types are capable
of regeneration. Nerve cells and germ cells produce factors that strongly inhibit
apoptosis (37), as removal in these cases has deleterious consequences. In case
of severe infection it can make sense to purge nerve cells (28).
Recent models dealing with purging-type phenomena have involved sto-
chastic models assuming finite populations. The key insight from the study of
antiredundancy is the ability of particulate, hierarchical systems to exploit cellu-
lar turnover to eliminate and replace deleterious components from populations.
3.5. Spatial Compartmentalization
Compartmental systems are those comprised of a finite number of macro-
scopic subsystems called compartments, each of which is well mixed. Com-
partments interact through the exchange of material (22). The spatial
compartmentalization of reactions leads to robustness by minimizing covariance
among reaction components participating in functionally unrelated processes.
Thus spatial de-correlation through compartmentalization substitutes for tempo-
ral correlation in biological functions. Robustness is achieved in at least two
ways: (1) minimizing interference—chemical, epistatic, or physiological, and
(2) minimizing mutual dependencies and thereby attenuating the propagation of
error through a system. The study of spatial compartmentalization is particularly
rich in theoretical ecology and epidemiology (34), where it has been used to
explore the maintenance of antigenic diversity, restrictions on pathogen viru-
lence, and seasonal forcing, and more recently in molecular biology, where pro-
teins have been found to be compartmentalized (48).
