Re: systems biology

[Tim Gwinn <***>]

Jerry,

I agree that systems biology *wants* to overcome reductionism. The question
is: *why* does it want to overcome reductionism? The only reason would be
that there are properties of a system-as-a-whole which are not captured by
analysis of the parts alone, and they want to capture those nonfractionable
"system" properties in models. This means that there are more ways to
analyze the whole system than simply as the inverse of the ways of composing
the system additively from its parts; that is, the number of analytic models
must be greater than the number of synthetic models.

In such a case (# of analytic models > # of synthetic models), then you have
a Rosennean complex system [LI 154, 176-177]. This is precisely what "Life
Itself" describes in detail. This also means that such a system has
noncomputable models.

Any computable model is a simple system. Any finite set of computable models
is still a simple system. Simple systems have analytic-models = inverses of
the synthetic-models. That is, systems biology approaches (such as
Genomatica) which rely only on computer-based models therefore will be
modeling simple systems. They cannot model a complex system, except locally
and temporarily. Most importantly, those "system" properties, which are lost
when the system is fractionated into parts, will not be captured in a finite
set of computable models. Because for a model to capture nonfractionable
properties of a system, it must be an analytic model which is not also a
synthetic model, since all synthetic models are by definition fractionable.
Therefore, the particular analytic model(s) which capture the
nonfractionable properties will be noncomputable ones, since all synthetic
models are computable. This doesn't mean that the computable models from
systems biology aren't useful for some purposes, but they are not capturing
the nonfractionable properties.

Now, if they have sets of models at one level (molecular biology) and
completely different sets of models at another level (system interactions),
where the 'parts' in the latter sets are not directly parts in the former
sets, then there is no way to logically reduce the latter to the former; so,
then have they "overcome" reductionism in that way? Well, it is not a
decided question, because again, the models at both levels are computable.
This means the *possibility* remains that they could reduce the latter
models to the former. Until one can characterize the relationships between
these levels of models we cannot say whether or not these disparate sets of
models can be collapsed into one larger computable model. This is in
contrast to a situation where the system interaction level possesses a
noncomputable model, which would automatically mean that those interactions
cannot be reduced without losing the nonfractionable properties.

Regards,
Tim


> -----Original Message-----
> From: ROSEN Forum [mailto:*** Behalf Of Jerry
> Zhu
> Sent: Tuesday, February 08, 2005 2:20 PM
> To: ***
> Subject: Re: systems biology
>
>
> Tim,
>
> It seems to me that it is clear that the systems
> biology is to overcome the reductionism in molecular
> biology by viewing the processes in a cell as systems
> of interactions much in similar way to view what's
> inside Boeing 777.
>
> See a company below doing systems biology and got lots
> of funding and investment. What do you anticipate that
> company?
>
> http://www.genomatica.com/index.shtml
>
> Jerry