Uncommon Descent Serving The Intelligent Design Community

Facilitated variation — the “conceptual completion” of Darwin’s theory of evolution?

arroba Email

Marc Kirschner’s idea of “bias toward useful variations” sure sounds teleological. I personally am biased to score well on standardized tests. And the reason for this bias is that I’m an intelligent agent who knows how to take these tests and score high. So perhaps Marc Kirschner’s theory of facilitated variation is a form of ID? You think he would warm to that idea? Probably not.

Answers to Darwin’s dilemma
Evolution is biased toward useful variations that emerge from genetic similarities shared by all living organisms
By Marc Kirschner
(June 9, 2006)

Biologist Marc W. Kirschner is founding chairman of the systems biology department at Harvard Medical School where he and his colleagues study the temporal and spatial cues that affect embryonic development. With co-author John Gerhart of the University of California, Berkeley, he explores natural selection in the face of biology’s most recent discoveries in The Plausibility of Life: Resolving Darwin’s Dilemma, which was reviewed by University of California, Irvine evolutionary biologist Francisco J. Ayala in the April issue of Science & Theology News.

With the release of the genome sequence in 2000, biologists finally realized that 22,500 is the magic number of genes needed to produce each person. That number is only one-and-a-half times that of a fruit fly and only six times the number in a bacterium — the simplest organism living on this planet. Looking at those genes, it was found that 15 percent were quite similar to those of bacteria, 50 percent were similar to those of a fruit fly, and 70 percent of flies were similar to those of a frog.

Humans are incredibly complex. Each person’s body contains about a hundred trillion cells and has probably thousands of cell types. So the question naturally arises: How do you get variety out of so few genes? How can animals be so different when many of the genes are the same? In short, where does novelty come from?

French naturalist Jean-Baptiste Lamarck had an answer to how diversity arises, especially in regard to the way variants become so well-suited to the conditions that exist in the environment. Lamarck thought that the environment caused the change. And this idea was so powerful that even though Darwin rejected it in The Origin of Species in 1859, 10 years later he embraced many of those same concepts. He agreed that the environment could in fact induce certain types of change. It took evolutionary biology many years to return to the ideas that Darwin had proposed in 1859 — that mutation is random, and that selection does not create variation but is merely a sieve that filters it.

This leaves a series of open questions: Is variation in anatomy and physiology, produced by random mutation, itself random? Or is it, in fact, biased? If it were biased, how could it be biased?

Our theory of facilitated variation is an effort to explain how animals convert random genetic variation into variation in structure and function that is biased to be most useful in evolution. Facilitated variation is not at all random. It doesn’t assume that need derives change, but it does explain why Lamarckian genetic variation produces nonrandom changes. Further, these changes are easily achieved with a rather small number of mutations, and, therefore, can happen quickly. They tend to be nonlethal because only nonlethal variation is inherited. And they tend to be appropriate to selective conditions. So even though mutation is random, change in the organism, physiology, anatomy and behavior can never be random. Change depends on what already exists.

There is a program on the Discovery Channel called “Monster Garage” in which host Jesse James shows how to convert a Ford Mustang into the world’s fastest lawn mower or a 1984 Porsche into a golf-ball retriever. But I doubt they could change a Cadillac into a violin. The point is that some changes are easy, while others are impossible.

What is easy? When you look into the processes, making a different-shaped limb, making a hand versus a flipper or a wing — that should be relatively easy to do. Making a larger brain, that should be relatively easy to do. Adapting to a different food supply, that should be rather easy to do. But the real insight that John Gerhart and I address in The Plausibility of Life: Resolving Darwin’s Dilemma is that on a cellular level, making new structures, even making a new eye, should not be that difficult either.

Constraints that empower

A key lesson from modern biology is that the animal is filled with genes and processes that do not change. Many of our genes are the same as a fruit fly’s. But these unchanging — and you might even say constrained — processes facilitate variation around them so that overall evolution is rapid.

I’ll give a couple of analogies. Imagine a monkey trying to write the word “monkey.” Give him a Bic pen, or even a Montblanc pen, and a piece of paper — he will produce only scratches on the paper. But give him a typewriter, and there will be a real possibility that he can pound out the requisite six letters in the right order in about 10 years. Not rapidly, but he could do it. The typewriter constrains the activity of the monkey to make letters, and letters have a chance to be useful. Pen scratches do not. If we constrain this monkey still further, using a computer program so that every time it punches a key it types a random English word out of the dictionary, it will probably produce the word “monkey” in less than a day — more constraint but more useful variation.

Let’s take a very common biological example, which is criticized by the people in the intelligent design community: the making of different kinds of vertebrate limbs. We have a great variety — we have our hands, we have bat wings, we have flippers, we have horse’s hooves. And if you think about it, making limbs seems rather difficult to achieve because not only do the bones need to be shaped in new ways but the muscles must also be appropriately attached to the bones in a new configuration. And that’s not enough. The nerve cells must be attached to the new positions where these muscles are.

That’s still not sufficient. A circulatory system that will appropriately nourish the bones must be generated, and the muscles and the nerve cells must be in their proper locations, along with a number of other simultaneous events.

It seems highly unlikely that this could happen until you look at the mechanisms by which the limbs were generated in the first place. When the bones are placed in some new location — and it can be done experimentally — the muscle cells will migrate out in an exploratory way from the flank of the body and take their position relative to the bones, wherever they are. As for nerve cells, nerve axons are produced in superfluous numbers, again exploring the periphery. If they attach to the right muscles, they will be stabilized and remain while those that cannot find muscles will die away and disappear.

The vascular system is another exploratory process that goes out into the periphery looking for regions with low levels of oxygen and, therefore, involves no preconceptions of where things are. If you look at the veins and the arteries in your right and left arms, which are genetically identical, you’ll see that they are quite different from each other because they have taken different random paths to explore the periphery.

Given that making a limb in the first place is much easier and changing a limb is much easier because the placement of the bones can be changed and the means by which the nerve cells attach to the muscles or the way the vascular system is generated does not have to be changed. These all adapt to the existing change.

Making new connections

A lot of evolution has to do with new connections between proteins and new connections between proteins and DNA. As it turns out, protein reactions can change easily and — in a way — that is quite useful.

As an analogy, imagine building a structure out of Lego blocks and building one out of modeling clay. Modeling clay is much more flexible, but with Lego blocks, which are each completely constrained, you can build the Eiffel Tower, or you can build a soccer ball. Although they are rigid, they have an easy connectability, which allows them to be arranged in new configurations. And that is similar to the way in which proteins and processes within the cell can easily interact in new configurations to generate new structures. They are rigid themselves, but they are “deconstrained” in the kinds of structures they can produce.

With the incredible numbers of cells and genes, there cannot possibly be genes or processes specifically for hands, brains or beaks. The processes that generate the structures of the body are more like carpenters or electricians or plumbers. They can build a beautiful church, or they can build a fast-food restaurant. They are versatile. And it is their versatility that makes it easy to understand change in evolution.

So the variety of tissues in our bodies is made largely from different combinations. The surprise in the last decade is that we now begin to understand what makes these things so combinable — the brilliance of a Lego block on a molecular level. Naturally, this ability to reuse different genes in new combinations, which is used within our bodies to generate new tissues, allows new uses in evolution.

I think that how molecular biology and developmental biology will help explain evolution and that will be the big story of the next decade in biology. But just as important is how evolution will help us understand ourselves.

Up until now, laboratory scientists have chosen to work on systems of limited variation — limited genetic variation, limited environmental variation. It’s easier to work in those systems. But in the future it will be the study of variation, whether normal or pathological, that will become the center of much biological study. Variation on a cellular level is critical to understanding human disease. We would dearly like to know how cancer cells vary so much and how they use their variations to evade the deadly poisons we throw at them.

Facilitated variation may have led to a conceptual completion of Darwin’s theory of evolution, but the subject of evolution and the study of evolutionary processes are far from complete. There’s a lot to learn. However, I hasten to add that it is only by testable scientific explanations that we will ever turn this understanding into something we can use for our own benefit.

This essay is adapted from “The Plausibility of Life: Resolving Darwin’s Dilemma,” remarks delivered by Marc W. Kirschner at the Cambridge Forum in Cambridge, Mass.

"The processes that generate the structures of the body are more like carpenters or electricians or plumbers.” This is all very well but we stopped believing in little men inside our TVs to make them work, when we turned five. Do they think there are the equivalent of little thinking agents "carpenters or electricians or plumbers" hiding inside our cells. Is there a blue print? Is it still thought to reside in the DNA or is there another heritable component that we need to discover? Is it all in the Junk DNA? Do the genes first make the "dumb" plumber machine that then reads the plumbing instructions and carries them out in programmed cooperation with the "dumb" electrician and "dumb" carpenter machines? Does this senario not cry out even louder for an intelligent architect? There is room for a revolution in thought here. idnet.com.au
Can we in ID predict what this certain something is? Does the meaning of the word genetics need to be changed? Go we need a new field of bio-organizationetics? What the common man means by inherited, and what we have been led to believe, is that what we are born as, is written in our DNA. Is this not true? idnet.com.au

"With the incredible numbers of cells and genes, there cannot possibly be genes or processes specifically for hands, brains or beaks. The processes that generate the structures of the body are more like carpenters or electricians or plumbers."

My female relatives have a presumably genetic defect that leads to a misshapen left little toe nail. This has been the phenotype for 4 generations. How can this be explained separate from their genes?

How can he be so confident that "there cannot possibly be genes or processes specifically for hands, brains or beaks."

I would say the opposite. There must be genes or at least higher order organising control programs for hands etc. Don't HOX genes work by causing these subroutines to repeat?

He's probably right that coding genes aren't responsible for many things. I doubt coding genes are responsible for instinctual behaviors, for instance, but instinctive behaviors are nonetheless heritable and are specified by *something* in the germ cells. -ds idnet.com.au
Re #10. In what sense is pentadactilism univeral? From Wikipedia: Tetradactyly is the condition of having four digits on a limb, as in amphibians and many birds. The hind limbs of dogs and cats also have only four digits. Tridactyly is the condition of having three digits on a limb, as in the Rhinoceros and ancestors of the horse such as Protohippus and Hipparion. These belong to the 'Perissodactyla'. Some birds also have three toes. Bidactyly or didactyly is the condition of having two digits on each limb, as in the Two-toed Sloth, Choloepus didactylus. In humans this name is used for an abnormality in which the middle digits are missing, leaving only the thumb and fifth finger. Cloven-hoofed mammals (such as deer, sheep and cattle - 'Artiodactyla') walk on two digits. Monodactyly is the condition of having a single digit on a limb, as in modern horses. These belong to the 'Perissodactyla'. ??? Mark Frank
If evolving change in limbs is so easy, why is pentadactilism so darn universal. There is no living species of quadruped with a polydactile limb as its prototype. Why? I dare you to explain it by NDE. Gould couldn't. Pentadactilism is universal because, only because, nature is determined to follow the design laid out for it. Convergence is so universal because, and only because nature is determined to follow the design laid out for it. Animals have this incredible ability to dynamically adapt to change, such as a different sized limb because they were masterfully designed. Lego was masterfully designed. bFast
Is it just me or is Kirschner just taking evidence for modular design and force-fitting it into the NDE narrative? I personally don't find these discoveries too surprising. Heck, I'm sitting at my desk right now working on a software engineering project that uses such concepts. I just wish my classes would automatically reconfigure themselves in order to provide new functionality as easily... Patrick
A couple points: Regarding "facilitate variation", this idea can be broken down more explicitly. First, to what degree is the shape of diversity shaped by the distribution of phenotypes that arise out of random mutation ? This distribution is certainly important since something that is more likely to occur via mutation, and has equal selective advantage relative to another mutation, would be expected to be more likely to fix. The fact that the probability distribution of available phenotypes plays a role in shaping diversity seems trivial. However, to what degree is the shape of this distribution "selectable" ? I am not sure if Kirschner is arguing that this distribution has been directly selected on over evolutionary time, rather than indirectly selected. It seems that he may be saying this distribution is itself a target of natural selection, but that is unclear to me. Nonetheless, whether the shape of this distribution has been selected for or not, he is arguing that it is important in understanding how evolution has occured. Re: 7. Facilitators are not in the same camp as designers in the sense of ID. Waves facilitate breaking rocks on the shore. bdelloid
Where there's facilitation there's a facilitator. The semantic games that chance worshippers play to avoid loaded terms connotating design never cease to amuse me. Of course they're never quite successful at it because any thinking person can't help but know that where there are machines there are machinists, where there are codes there are coders, and where there is facilitation there are facilitators. :roll: DaveScot
Re #2. I assume that in this context "useful" means useful to the organism in helping it reproduce. It is the same sense that we can ask "what's it for" for any living system. Facilitated variation could achieve these ends by a mechanism that imposes constraints so that mutations which lead to lego bricks are more likely than mutations that lead to meaningless goo. I am no geneticist, but it is striking how mutuations can lead to developments such an additional arm which are constructing a meaningful unit of function (even if it is not appropriate in this case). Re #4. I think there are two uses of the words "random" here. Random meaning "all mutations are equally likely" and random meaning "not directed towards an end". I am guessing that Kirshner is saying the first is not true but the second holds. Mark Frank
Kirschner and Gerhart admit that novelty and complexity are inherent in useful mutational changes, and that environmental adaptations through natural selection are not enough. This novel approach to evolutionary change is supposed to be an answer to the design inference proposition, and an explanation for the development of upward complexity. They talk about things like 'evolutionary novelty', 'coordination of development, and 'exploratory behavior' during embryonic development. These mechanisms are supposed to differ from intelligent design in that they are modifications done 'on the fly', and thus more flexible than a rigid pre-designed structure. But wait! A possible purport of this theory is that it negates the need for a designer. I see it as similar to NS/RM, but supposedly consisting of a 'smarter' evolutionary process. But does it/can it really fulfill the evolutionist's dream? If it does, it may in fact be the answer to Darwin's dilemma. I just called around and found a copy of the book at Border's in Tempe AZ, and I'll pick it up this afternoon. After reading it, I may have a further comment. This article by Daniel Hartl (12/05 Harvard Magazine) also sheds some light. http://www.harvard-magazine.com/on-line/110512.html leebowman
Kirshner seems confused about the word "random". Here he says: "Facilitated variation is not at all random." and "Is variation in anatomy and physiology, produced by random mutation, itself random? Or is it, in fact, biased? If it were biased, how could it be biased?" What is he talking about ? He says that mutations are random, but the phenotypic results are not ? I think he is simply trying to convey an idea that the phenotypic output from a random mutation is not what one (ie, he) would expect - that they have a greater chance of producing a positive selective result than would be expected. But, what is the "expected distribution of phenotypes" that arises out of random mutation and how is this different than the "observed distribution of phenotypes" ? In the end, I have no idea what he means when he says the distribution of phenotypes is not random. Of course it is random - it just may be a different probability distribution than he would expect. Perhaps the mean of the phenotypic advantage is higher than he would expect, but that isn't to say it is no longer a random distribution. Does Kirshner think an unfair coin, with a probability of heads being 90 %, doesn't produce a "random distribution" ? Also, for another take on this, see this article by Brian Charlesworth: http://www.sciencemag.org/cgi/content/full/310/5754/1619 bdelloid
They tend to be nonlethal because only nonlethal variation is inherited.
Erm. These variations tend to be non-lethal because they must be non-lethal of necessity?
The typewriter constrains the activity of the monkey to make letters, and letters have a chance to be useful. Pen scratches do not.
And yet, in the evolution of technology, pen scratches preceeded typewriters. Think of how much more variation is available for selection to act upon when a pen is used, rather than a typewriter. And think of just how restrictive a typewriter really is, as in contrained (typically) to a single alphabet. Isn't this just another teleological probability argument? Mung
"Useful" is an inherently teleological notion. We're never just talking about usefulness as such but usefulness for some end. What ends are we talking about? And how exactly is facilitated variation biased to achieve those ends? This, it seems to me, is a gaping hole in Kirschner's approach. William Dembski
I assume it means what it says - the idea that there is a bias towards useful variations. There is no reason why this should not itself evolve. Anything that constrains variation in ways that limits destructive variation and enables constructive variation is going to provide a rather large competitive advantage. Mark Frank

Leave a Reply