While catching up my reading list, I finally got back to Stephen’s recommended paper titled “Mitotic-Exit Control as an Evolved Complex System.” (Mitotic-Exit Control as an Evolved Complex System: Cell) Concepts from this paper seem to capture some of the thoughts that we’ve been addressing in this discussion.
An informing quote comes from the first section titled “Top-Down Design vs Evolutionary Engineering.” It says, “Simple systems, even complicated ones, can be decomposed into modules or pieces at all scales. An automobile or modern jet aircraft, as complicated but noncomplex systems, can be understood as the sum of their subsystems: the computers, the engine, braking systems, the flight stabilizers, and other major subsystems all have a clear function in the whole. Each of these can also be broken down and understood in terms of yet smaller components, down to the most basic mechanical and electronic parts. Viewed from the design perspective, this complicated system can be put together by many engineers, each working independently on separate components according to a master, top-down design plan. Traditional engineering design depends on this top-down, modular approach and the decomposability of the system. The system must perform precisely as the sum of all the components: it is designed under this assumption.”
It then says, “Complex systems generally cannot be decomposed and built this way. Attempts to do so have met with spectacular failure.”
I completely agree, although with one exception, and that’s when Jesus Christ is the designer. Rumor is that he’s an exceptional design engineer. I would also expect that he uses a “top-down design” for the Cell since it’s not so complex that he can’t do it. I’ll even bet that he gets it right on the first pass.
The paper goes on to propose that “evolutionary engineering” can achieve what the “top-down design” approach cannot. While it’s recognized that complicated systems (the paper distinguished between a complicated and a complex system, but doesn’t identify the threshold where one ends and the other begins) can be decomposed into it’s component parts, the paper suggests/declares that complex systems exhibit function that is more than the sum of its parts, and concludes that “Complex systems generally cannot be decomposed and built this way.” The paper illustrates its point by contrasting the failed US Air Traffic Control System back in the 1990’s with the global internet of today that has incrementally grown as many different users and developers have constructively added capability over time. New capabilities that were never envisioned by the original architects are showing up daily. Thus, “evolutionary engineering” is the best approach for complex systems. (I don’t agree.)
The FAA system is supposedly an example of how a complex system employing a “top-down” design (systems engineering at its best?) has limits especially with “emergent” capabilities that are beneficial but not foreseeable. The global internet in contrast is used to illustrate “evolutionary design” that grows not by grand design, but by incremental small local improvements that improve the overall system thru natural selection. If a local improvement fails to improve a sufficiently large segment of the internet so that its cost of development and operations exceeds its value to the larger system, it will die. The reason is that it won’t generate sufficient revenue to its developers/operators to generate a profit. It will be removed by its developer(s) at the request of its investors to save money. Thus, the global internet earns its label as an “evolutionary design,” at least at a superficial level.
But let’s look a little closer. Things may not be as they appear. The contrast between a “top-down design” and “evolutionary engineering” has some hidden assumptions that may not be obvious on a first reading. In the FAA illustration, the designers are fallible human beings who don’t do well with complexity. They are easily confused, loose track of important details, lose sight of the forest for the trees, cheat, lie, and steal, and generally are inept when dealing with complex designs. Most importantly, they can’t anticipate “emergence” or “synergy” easily. Thus, if humans are the designers, it is easy to conclude that “top down” designs can’t explain complex systems, especially for the Cell.
If, however, the designer is Jesus Christ, we may need to reassess our conclusion about “top down” designs. When I get to heaven, my first request is to visit the Lord’s technical library where He keeps all His design documents. I want to see the document titled “First Eukaryotic Cell.” Then, I want to see the Change Documentation for the Fall, i.e., when Adam and Eve ate the apple, how did He change the design? I assume also that he has an amazing video library. I’ll want to watch the creation in Gen 1, the Flood, the Exodus, etc. The list of documents and videos I will want to see is endless. It may be a little weird, but that’s my idea of heaven, viz., to review engineering documentation. (Yes, I need therapy!) Can you imagine my disappointment if He says, “Sorry, no design; I just put together some solids, fluids and gases, tossed in few rules, shook it and came back in a few billion years to see if anything came out of it.” To his surprise, he finds a bunch of humans drinking beer and arguing over where they came from!
On the “evolutionary engineering” side, when we look closely, we see an implicit assumption that is questionable. The example, the global internet, may not be a good analogy for the Cell for an important reason. Natural selection for the global internet works at the level of App that can be connected to the internet easily with no potential for harmful effects. There are many changes being introduced daily in the sense that any user can develop and insert a new capability at any time without coordination with other users. The “random change” element of the “evolutionary engineering” model finds its home in a very simple and non-destructive activity, viz., develop an App, connect it to the internet, let the search engines find and expose it, and let social media praise or kill it. The “natural selection” component is provided by the market place filtering out the unfit/unprofitable. On the surface, it sounds like evolution in good form. But is it a good analogy? I say no. Let me explain.
What’s the difference between the FAA systems and the global internet that should make us take a step back and reflect a bit? The component parts of the FAA system follows the description of a complex system identified by INCOSE , viz., “systems with rich interdependence among diverse components, non-linearity, open systems boundaries, networks of causality and influence (vice linear causal chains), emergence, varied and changing system goals, self-organization, and multi-level adaptation.” I need to emphasize in particular the attribute of “rich interdependence.” Does this apply to the global internet? For the internet developers of new Apps can develop independently, publish a link so that others can find it, and sit back and watch market forces go to work. But what if there’s extensive interdependence between every new App and many of the internal functions of the internet? If that were so, then every App developer would have to coordinate with the global internet staff of function managers to determine if their App will indeed connect without being destructive. This, of course, is the bane of “top down designs.” We wouldn’t have the internet today if that were the case.
Now we need to ask the question: “Which is the better model for the Cell?” Can we say that every new protein or group of proteins coming from a random gene/allele mutation is like adding an App to the internet? If so, we’d have to admit that the cell is virtually insensitive to the new resulting proteins coming from those mutations. Can we say that? If not, then the global internet analogy as a complex system exhibiting “evolutionary engineering” is not a good one.
This takes us back to the question of the which applies best to the Cell, does it reflect a “top down design” or “evolutionary engineering?” The answer seems to hinge on the sensitivity of all the Cell’s proteins to a new one coming from a mutation. So far you guys claim that the Cell is more like the internet analogy where changes to a protein are not destructive. That, of course, implies that you don’t see “rich interdependence” between the Cell’s proteins. Of course, you claim that RFR overcomes the potential destructive effects of mutated proteins. If that’s the case you have good evidence to support it, right? How many proteins in the Cell have been observed, measured, and published to demonstrate this insensitivity (by RFR). If only a few, is it fair to claim its universal in the Cell? If that were the case, I wonder why Doug Axe’s paper (JMB-2004-341, 1295-1315) shows the Cell to be very sensitive? If I recall correctly, as he mutated a specific protein, it became non-functional very quickly. From that he inferred an estimate of the prevalence of functional protein sequences at the level of one in 10^77.
Maybe I’m over simplifying, but your apparent assumption that the Cell’s proteins are insensitive to changes in other proteins resulting from mutations of nucleotide bases in the genes/alleles is not a good one. If indeed it is a bad assumption and the Cell exhibits the kind of complexity described by INCOSE (rich interdependence among diverse components), then “evolutionary engineering” cannot happen.
Hopefully, we’re not back where we started. I saw the internet analogy as a really bad one, so I decided to try and say why and see what you think. I’m sure you’ll push back and claim that protein interdependence is there, but because of RFR the destructive effects are eliminated. Thus, the “evolutionary engineering” and its internet analogy are legitimate. Perhaps, but if the interdependence came first before RFR, how did the mutations not destroy the Cell long before RFR appear? Mutations work against RFR ever coming to the rescue. Of course, if you reject interdependence of the proteins, then the internet analogy works. But then you wouldn’t need RFR, and it would never evolve.
What say you?