Evolution Icon Evolution

The Fairyland of Evolutionary Modeling

Fairy_passage.jpg

Evolution takes too long, so we’ll just speed it up in a computer. That’s what two Darwinians scientists from the University of Barcelona decided to do. They created maps of virtual worlds where new adapted animals emerge. But even they, with all their creative modeling, couldn’t get natural selection to bear the load put on it.

Models are useful in science. They’ve been used since Victorian times. William Thomson (Lord Kelvin) said that unless he can build a model of something he can’t really understand it. The models of 19th- and 20th-century scientists have usually been testable against the real world. Observation provides important tethers to models, so that they don’t fly off into fairyland.

Where, though, are the connections to reality in a paper by Isaac Salazar-Ciudad and Miquel Marin-Riera, published in Nature? One will look in vain for any mention of fossils, mammals, arthropods, or any other animal. The nearest natural thing they discuss is a tooth — but even that is an idealized, virtual tooth in their dreamed up world of visions and simulations.

The University of Barcelona describes their approach: “3D simulation shows how form of complex organs evolves by natural selection.” Notice that they call theirs the first simulation of its kind:

Researchers at the Institute of Biotechnology at the Helsinki University and the UAB have developed the first three-dimensional simulation of the evolution of morphology by integrating the mechanisms of genetic regulation that take place during embryo development. The study, published in Nature, highlights the real complexity of the genetic interactions that lead to adult organisms’ phenotypes (physical forms), helps to explain how natural selection influences body form and leads towards much more realistic virtual experiments on evolution. (Emphasis added.)

There’s a conundrum for you: “more realistic virtual experiments.” Try this exercise: find the realism in their model. We read about “virtual simulation of evolution” and “models of virtual evolution by natural selection of form.” Here’s a gem: “Evolution takes place virtually on the computer in populations of individuals in which each individual can mutate its genes, just as this works in nature.” Nature itself nowhere appears, though, except in the name of the journal that published their imaginative excursion.

Keep in mind that the Darwinian evolutionists’ burden is to demonstrate that natural selection actually produced novel, innovative structures, the classic examples being an eye or a wing. It’s not convincing to look at existing eyes or wings and simply assert that they evolved by unguided processes. Nor is it convincing to look at existing genes and make that assertion. Skeptics of neo-Darwinism will suggest other explanations for the observations, such as genetic drift or intelligent design.

Salazar and Marin-Riera seem to know this. That’s why they make a glaring admission in explaining why they couldn’t be realistic, but had to go virtual:

“Right now we have a lot of information on what changes in what genes cause what changes in form. But all this is merely descriptive. The issue is to understand the biological logic that determines which changes in form come from which changes in genes and how this can change the body”, explains Isaac Salazar, a researcher at the University of Helsinki and in the Department of Genetics and Microbiology of the UAB, and lead author of the article. In nature this is determined by embryo development, during the life of each organism, and by evolution through natural selection, for each population and species.

But in the field of evolution of organisms it is practically impossible to set up experiments, given the long timescale these phenomena operate on. This means that there are still open debates, with hypotheses that are hard to prove experimentally. This difficulty is compensated for by the use of theoretical models to integrate in detail the existing experimental data, thus creating a virtual simulation of evolution.

But their brief reference to “existing experimental data,” remember, does not demonstrate the validity of natural selection, because alternative explanations exist. It’s up to them to prove it. Otherwise, their “virtual simulation of evolution” floats free in the realm of imagination. Their whole paper is built on “theoretical models” they wrote and that other Darwinian evolutionists wrote. The real-world constraints on their theoretical model are just other theoretical models!

Even so, it’s astonishing to read on and find out that with all the freedom in their imaginary world, they still couldn’t get natural selection to accomplish much. We are offered three “visions” of evolution (an instructive choice of words). The first “vision” was that any change was adaptive. That one got shot down, leaving two others with little for natural selection to do:

This simulation enables a comparison of the different hypotheses in the field of evolution regarding which aspects of morphology evolve most easily. The first vision is that all metric aspects of form contribute to adaptation and that, consequently, all are fine-tuned by evolution over time. The second vision is that some aspects of form have greater adaptive value and that the remainder evolve collaterally from changes in these. The third is that no aspect of form is intrinsically more important, but what is important adaptively is a complex measurement of the form’s roughness.
“What we have found is that the first hypothesis is not possible and that the second is possible in some cases. Even if ecology favoured this type of selection (the first vision), embryo development and the relationship between genetic and morphological variation imposed by this is too complex for every aspect of morphology to have been fine-tuned. In one way, what we are seeing is that natural selection is constantly modelling body forms, but these are still a long way from perfection in many ways“, points out Salazar.

It must be obvious that perfection in a virtual world is a lot different from perfection in the real world. Watching Illustra Media’s film Metamorphosis, about butterflies, it would be hard to imagine creatures gifted with better engineering and artistry. Salazar and his companion focused on virtual teeth, using their own subjective criteria — nothing from the real world.

By “roughness,” they are referring to another theoretical construct in evolutionism, the “fitness landscape” imagined by Sewall Wright. The fitness landscape has peaks and valleys. An organism might attain a local optimum fitness but get stuck there, with no way to climb to a higher peak without having to decrease its fitness first. Interesting as the fitness landscape concept is, it’s incidental. What matters is that Salazar’s entire evolutionary model is adrift in virtual space — and even there, natural selection doesn’t do much.

Remember, this was not some backwoods thought experiment, but was published in the leading science journal in the world, Nature. And it got good press. In the same issue of Nature, P. David Polly gives his perspective on the paper, under the headline, “Evolution: Stuck between the teeth.” Here’s the upshot:

A computer model of tooth evolution designed to assess the impact of developmental dynamics on natural selection reveals that complexity reduces the likelihood of maximum fitness being attained.

Polly was not being critical. He congratulated their “groundbreakingly realistic computational model” of how natural selection works. (The realism still eludes us.) Polly spends some time discussing how real teeth develop in living animals, but that’s not the issue. Remember, the burden of the neo-Darwinist is to explain how unguided natural processes achieved the high degrees of adaptation that we actually observe in the living world.

Polly’s digression into real teeth doesn’t help, therefore, because he admits that “the authors created artificial populations of developing teeth that they subjected to mutation and selection.” It’s all simulation that cannot be tested in the real world. Remember? It takes too long.

When we see what the authors did, we have to sigh: “There you go again.” They snuck information into their evolutionary algorithm. That’s the only way they could get evolution to (at least partially) work. Look at all the free parameters in their arbitrary scheme:

Fitness was assigned to each individual on the basis of either its phenotypic or its functional similarity to an arbitrarily chosen target phenotype. The latter can be thought of as the phenotype that conveys the greatest fitness in the ecological context of the simulation — a peak on the adaptive landscape, as described by the geneticist Sewall Wright. The fittest individuals were then selected as the parents of the next generation and the simulation was repeated.

But even with all the coaching of their imaginary organisms, they could only get some of the virtual individuals to reach their full fitness potential. Alas:

Salazar-Ciudad and Marin-Riera have shown that not only are suboptimal dead ends an evolutionary possibility, but they are also exceedingly likely to occur in real, developmentally complex structures when fitness is determined by the exact form of the phenotype.

Well, they couldn’t leave natural selection powerless to create endless forms most beautiful, so they tweaked the model to steer it up the fitness peak:

However, the authors also found that when fitness was determined by functional properties instead of the phenotype itself, the adaptive peak was usually reached. This is because many different phenotypes can have the same functional properties — a herbivorous mammal, for example, simply needs grinding and chewing surfaces on its teeth, regardless of how the surfaces are constructed. Thus, there are many more paths to a functional adaptive peak than to a phenotypic one, especially for a phenotype that has a complex developmental system, such as a tooth.

Of course the computer organisms will reach the adaptive peak when you define it in terms of functional properties that you yourself reward your virtual organisms to find, by calling them the fittest and only allowing them to reproduce. Try it with no rewards, and no guidance (as neo-Darwinism is supposed to work), and the algorithm reduces to blind search.
Polly is impressed by the work, but ends with a series of questions that reveal the Darwinian mechanism as a series of promissory notes with no deadlines. Pay special notice to his reference to “knowledge gaps”:

The authors’ results highlight interesting questions. If phenotypic evolution is likely to get stuck on suboptimal adaptive peaks, what happens to the evolving populations? Do they simply persist in a suboptimal state? Do their ecological relationships change to fit their phenotype, thereby creating new adaptive peaks? Or do they become extinct as they are out-competed by populations that can reach a more optimal phenotype? Salazar-Ciudad and Marin-Riera’s focus on teeth — which can be studied empirically in living populations, among distantly related clades and in the fossil record — offers considerable potential for testing their evolutionary predictions and closing the knowledge gaps between genetics, development and macroevolution.

So they can be tested (implying they have not been), provided that one begs the question of whether unguided natural processes are responsible. Some day we’ll close those knowledge gaps! Just keep the funding on tap to pay our computer programmers. Meanwhile we see that Darwinian theory, which is sold to the public as so obvious in nature it should be the only one permitted in schools, works only in virtual space. And even there, it doesn’t work very well.

Image: Fairies Looking Through a Gothic Arch, John Anster Fitzgerald/Wikicommons.

Evolution News

Evolution News & Science Today (EN) provides original reporting and analysis about evolution, neuroscience, bioethics, intelligent design and other science-related issues, including breaking news about scientific research. It also covers the impact of science on culture and conflicts over free speech and academic freedom in science. Finally, it fact-checks and critiques media coverage of scientific issues.

Share

Tags

__k-reviewComputational SciencesNature