|
Author
|
Topic: Feedback and Control in Biology
|
Columbo
Member
Member # 113
|
posted 08. April 2002 00:52
The following diagram depicts in schematic form, a typical feedback and control system:
In a humanly engineered feedback and control system, a controller modulates the output of a system (the Plant) by sensing the plant’s output (Y), and comparing it to a control input (R). The difference (e), called the ‘error’ signal, is then processed by the controller so that the plant input signal ‘u’ minimizes ‘e’.
It is a commonplace that biological systems are replete with feedback and control systems. As an example, the following website at Yale, outlines the feedback and control mechanism involved in the human respiratory system:
http://www.med.yale.edu/caim/bme350/lec981007.html
As an engineer, I am curious about the origin and structure of feedback and control systems in biology. There are features of humanly engineered FC systems that seem to me to be teleological by definition, and so pose a problem when it comes to explaining not only their origin, but their very presence in biological organisms. For instance, the input signal (R) is an arbitrary reference value, chosen for a desired output. It would seem that in a biological system, it (R) would have to arise independent of any environmental factors. Another issue is of course, the fact that such systems are irreducibly complex. Finally, why are they so ubiquitously present? Wouldn’t a Darwinian paradigm predict open-loop systems, rather than closed-loop, as the more common feature?
I wrote to Dr. Michael Behe, to ask him if he would identify such a system for me at the sub-cell level. He suggested “the enzyme aspartate transcarbamoylase, which synthesizes the first compound in a pathway that leads to a class of molecules called pyrimidines. There are multiple steps in between the step catalyzed by ATCase and the final product. The final product, cytidine triphosphate, if its concentration is high enough, can then bind to ATCase and turn down its activity.” When I asked him what the control signal would be, he suggested that the binding affinity of ATCase for cytidine triphosphate might be a candidate.
Would anyone care to help me understand the nature of biological feedback and control systems, and how they might be accounted for in Darwinian, and in ID terms?
Columbo
[ 08 April 2002, 09:14: Message edited by: Moderator ]
IP: Logged
|
|
James A. Barham
Member
Member # 50
|
posted 08. April 2002 08:07
Columbo:
In my view, we have to remember that every model has its limitations. While the cybernetic feedback control model is very helpful for understanding certain aspects about living things, as you point out, it presupposes an overall teleological organization, and so cannot explain it. (Natural selection does the same thing, and so is no help, either.)
There are a few of us who are arguing that the solution is to come at the problem from the direction of nonlinear science and condensed-matter physics. For example, on F.E. Yates's "homeodynamic" model, the organism is modeled as a network of highly nonlinear oscillators, with inherent dynamical properties which collectively give rise to intelligent action, thus answering your objection.
This approach is better known at the level of the brain (esp. the work of Walter J. Freeman), but if Yates is right, it has quite general application.
See F.E. Yates, "Order and Complexity in Dynamical Systems: Homeodynamics as a Generalized Mechanics for Biology," Mathematical and Computer Modelling, 1994, 19: 49--74.
I discuss some of the philosophical implications of Yates's ideas in my own papers (why don't you contact me personally, if you would like more informaiton).
IP: Logged
|
|
Janitor@MIT
Member
Member # 125
|
posted 09. April 2002 15:52
Columbo—“Wouldn’t a Darwinian paradigm predict open-loop systems, rather than closed-loop, as the more common feature?”
Cornish-Bowden, Athel & Cardenas, Maria Luz (eds.). 1990. Control of Metabolic Processes: Proceedings of a NATO Advanced Research Workshop on Control of Metabolic Processes. Plenum Press. NY. pp. 159-172. (Accessed online 12/01/01.)
This section on interconvertable enzyme cascades is very interesting. Cornish-Bowden and Cardenas ask exactly the same question while tracing out these networks: why does the MAP cascade, e.g., consist of multiple kinases when one would suffice to do the job?
“… A cascade with a single cycle can generate responses with extremely high Hill coefficients (about 800) if the three conditions for high sensitivity are satisfied: (i) modifier enzymes near saturation; (ii) catalytic rather than specific effects; and (iii) inhibition stronger than activation. Furthermore, analysis according to the methods of metabolic control analysis give very high response coefficients (SZEDLACSEK, S. E., CÁRDENAS, M. L. and CORNISH-BOWDEN, A. 1992. "Response coefficients of interconvertible enzyme cascades towards effectors that act on one or both modifier enzymes" Eur. J. Biochem. 204, 807--813.)… However, this leaves unanswered the essential question of why three cycles are needed when evolution could easily have selected kinetic parameters and enzyme concentrations in the MAP kinase cascade to provide the same modest degree of sensitivity with just one cycle. Why three?” Maybe that’s a “Why ask why?” sort of question, but the authors conclude that the system evolved to modulate, amplify, attenuate control signals through stunningly complicated metabolic networks. They also note that this isn’t simply a matter of control over isolable pathways, but a problem of systemic integration. The concept of distributed control occurs naturally in this field, challenging the genocentric, and chemokinetic views. The existence of such a distributed control network must be difficult to account for on “numerous, slight, successive modifications.” You might want to take a look at “Snapshots of Systems—Metabolic Control Analysis and Biotechnology in the Post-Genomic Era,” by Douglas B. Kell and Pedro Mendes (Accessed online 12/16/01.) where the authors simulate modifying even a comparatively simple such system. Maybe there is someone out there who has some experience in reconfiguring/retraining complex distributed control networks?
IP: Logged
|
|
Columbo
Member
Member # 113
|
posted 10. April 2002 01:31
James: Thanks for the info, and for the invite to contact you. I will be traveling for a couple of days, so when I get back, I will try to follow-up expeditiously.
Janitor: Your reply looks very interesting, though I will certainly have to consult a good dictionary to absorb it all.
Thanks to both of you for illuminating what to me is an intriguing problem. More when I get back!
Columbo
IP: Logged
|
|
|
Printer-friendly view of this topic