Biology Reference
In-Depth Information
complexity of the molecular environment in cells in our theories. How much
all these details matter is controversial. In complex multilevel systems, it is
inconceivable that all the properties at one level have equally strong influences
over the properties and behaviour at higher levels, but even this presupposes
that levels or modules can be identified as discrete entities on the basis of time
or distance scales or of some other measure of the clustering of the components
involved. The paradox is that the only way to be certain that a simplified model
is an adequate representation of the properties of a more complex system is
to have a more detailed, complex (theoretical or experimental) model whose
properties can be compared. But usually, the reason for making a simpler model
is that the more complex model is not tractable in the first place. This may offer
a role for simulation, which will be considered at the end of this chapter, but
first there is a more fundamental question to be addressed.
4. SHOULD WE EXPECT METABOLISM TO BE
UNDERSTANDABLE?
As mentioned above, function plays a key role in biological explanations. To
some extent this is a legacy of religious and anthropocentric belief systems, but
the theory of natural selection implies that unless a component of an organism
is truly vestigial, it makes some contribution to the fitness of an organism and
it is this that constitutes its function. At a high enough level of organisation in
the organism, this is unlikely to be difficult to discern or controversial. No one
is likely to argue against locomotion as a function of legs, or gas exchange as a
function of lungs. It is how these higher level functions arise from the activities of
lower level components that is more problematic, and this has been increasingly
evident as techniques have been developed for targeted gene knockouts. In every
genome that has been sequenced to date, there is a substantial proportion of
putative genes for which no biological activity is known. This may start at
over 60% of the genes for a newly sequenced organism, and falls with time,
but it is difficult to get this down even to 50%. One technique for identifying
possible functions of gene products is to selectively knock out each gene in the
organism in turn and determine the effect on the phenotype. This has been done
systematically for the yeast
Saccharomyces cerevisiae
in the Eurofan project
(Dujon, 1998), but a large fraction of the knockouts show no overt phenotype.
There are two main explanations for the roles of these genes: either they only
prove necessary under environmental circumstances that have not been examined
in the laboratory or the difference that they make is too small to be detected
by normal laboratory experiments. This was tested by Thatcher et al. (1998),
who took a random selection of knockouts with no overt phenotype and grew
them in competition with the parent wild-type strain in a chemostat. In the
