Causality, Complexity and Computers
19
therefore very much 'in the system'. A proper understanding of the system
must therefore take account of these phenomena and explain rather than
ignore them.
So we come to the somewhat unsatisfactory conclusion that there are
no general rules for defining the limits of a system. The analysis has,
however, pointed the way we must go. It is one of our aims to give a
quantitative description of enzyme systems of this kind and in the following
we shall describe how we have approached the problem.
The Sensitivity Coefficient
We are faced with the situation that we have fairly detailed knowledge
of the structure of enzyme systems as well as information on the properties
and relations of the individual (isolated) enzymes and their substrates. But
since there are many interactions-sequential, competitive, inhibitory and
activating,-we must devise a method of studying each enzyme when it is
acting and interacting in the whole system.
We have already pointed out that the total elimination of enzyme
activity (by e.g. the classical genetic block) usually leads to elimination of
the measured property and sometimes to the elimination of the whole
organism. This kind of experiment therefore yields mainly structural
information and gives no quantitative answers as to the role of the enzyme.
(Fig. 6a). There is, however, a related experiment which immediately gives
quantitative information. This consists of 'modulating' the activity of one
enzyme and measuring the result on as many systemic properties such as
products, fluxes (and other enzymes!) as we may choose. We can e.g.
introduce into an organism a mutant enzyme with altered activity (as in
the strains in Fig. 5) or we could alter the quantity of enzyme by controlling
the nuclear dose, or we could inhibit it in some specific manner.
Let us consider the investigation of, say, a flux F through a pathway
in a system consisting of many enzymes. We shall assume (and may have
evidence) that the system is in steady state. I shall not here go into the
technical difficulties (which are not small) of measuring the flux of molecules in a particular part of an organism but shall assume that we have some
means of doing so. Similarly I shall assume that we can estimate the quantity
or activity of a chosen enzyme E. If we now compare, say, a different mutant
with enzyme activity altered by a step Ll E we can ask what change in the
flux, ;1 F, has resulted. (Fig. 6 b). It is best to express these as percentage
LIE
LIP
changes or as fractional changes, i.e. as E and F since this eliminates
scale effects. As the flux change depends on the magnitude of the enzyme
LIP LIE
change, the ratio, R, of these two quantities, i.e. F / E = R, is some
2*
19
therefore very much 'in the system'. A proper understanding of the system
must therefore take account of these phenomena and explain rather than
ignore them.
So we come to the somewhat unsatisfactory conclusion that there are
no general rules for defining the limits of a system. The analysis has,
however, pointed the way we must go. It is one of our aims to give a
quantitative description of enzyme systems of this kind and in the following
we shall describe how we have approached the problem.
The Sensitivity Coefficient
We are faced with the situation that we have fairly detailed knowledge
of the structure of enzyme systems as well as information on the properties
and relations of the individual (isolated) enzymes and their substrates. But
since there are many interactions-sequential, competitive, inhibitory and
activating,-we must devise a method of studying each enzyme when it is
acting and interacting in the whole system.
We have already pointed out that the total elimination of enzyme
activity (by e.g. the classical genetic block) usually leads to elimination of
the measured property and sometimes to the elimination of the whole
organism. This kind of experiment therefore yields mainly structural
information and gives no quantitative answers as to the role of the enzyme.
(Fig. 6a). There is, however, a related experiment which immediately gives
quantitative information. This consists of 'modulating' the activity of one
enzyme and measuring the result on as many systemic properties such as
products, fluxes (and other enzymes!) as we may choose. We can e.g.
introduce into an organism a mutant enzyme with altered activity (as in
the strains in Fig. 5) or we could alter the quantity of enzyme by controlling
the nuclear dose, or we could inhibit it in some specific manner.
Let us consider the investigation of, say, a flux F through a pathway
in a system consisting of many enzymes. We shall assume (and may have
evidence) that the system is in steady state. I shall not here go into the
technical difficulties (which are not small) of measuring the flux of molecules in a particular part of an organism but shall assume that we have some
means of doing so. Similarly I shall assume that we can estimate the quantity
or activity of a chosen enzyme E. If we now compare, say, a different mutant
with enzyme activity altered by a step Ll E we can ask what change in the
flux, ;1 F, has resulted. (Fig. 6 b). It is best to express these as percentage
LIE
LIP
changes or as fractional changes, i.e. as E and F since this eliminates
scale effects. As the flux change depends on the magnitude of the enzyme
LIP LIE
change, the ratio, R, of these two quantities, i.e. F / E = R, is some
2*