The science editor of HuffPo keeps asking me to come aboard and contribute a column, and I keep saying that I’m not going to write for free for a profit-making venture (their columnists are all wage slaves, or rather no-wage slaves, since they’re never paid). And I always take the opportunity to offer some “constructive” criticism about the page. One of my complaints is that they give a venue for James Shapiro, a colleague in biology at the University of Chicago, to attack neo-Darwinian evolution for completely bogus reasons. There is no scientific vetting of whether what Shapiro (or anyone else) says is correct. And in his case, his criticisms of modern evolutionary theory are dead wrong. So HuffPo perpetuates the idea that something is wrong with evolutionary theory.
Shaprio has four beefs about neo-Darwinism (the modern theory of evolution)
1. It’s not “gradual,” but can happen nearly instantaneously due to things like horizontal gene transfer, gene duplication, and other “self-engineering” features of the genome. In my post linked to above, I explain why this is fallacious. Some evolution can happen quickly due to horizontal gene transfer, but that’s not the main engine of adaptive evolution, and even gene duplication takes a long time for the duplicated genes to evolve new function. Also, groups of genes transferred “horizontally,” between very unrelated organisms, will persist only if they have a salutary effect on the fitness of the recipient.
2. Conventional evolutionary mechanisms (presumably natural selection and genetic drift) aren’t efficacious in creating big organismal changes. In this respect Shapiro comes very close to creationism. In fact, in one of his posts he explicitly uses a creationist trope:
The first problem with selection as the source of diversity is that selection by humans, the subject of Darwin’s opening chapter, modifies existing traits but does not produce new traits or new species. Dogs may vary widely as a result of selective breeding, but they always remain dogs.
Given the fossil evidence of transitional forms—showing that fish became amphibians, amphibians became reptiles, reptiles became mammals as well as birds, even-toed terrestrial mammals became whales, and early primates became humans (please, cladists, keep your objections to yourself!)—such a statement is simply embarrassing, and is identical to ones you’ll see in the creationist literature. Shapiro should know better.
3. Mutations aren’t “random” as neo-Darwinians contend. (By “random,” evolutionists mean “mutations occur regardless of whether they’d enhance the fitness of the organism.”) In fact, we know of no evidence for mutations occurring nonrandomly or “adaptively”, i.e., that the occurrence of mutations is somehow biased in a direction that makes them more likely to be favorable when they arise, particularly when the environment changes in a way that requires favorable mutations to fuel adaptive evolution. There has been some controversy about the occurrence of “adaptive mutation” in bacteria, but that’s died out because there’s simply no evidence that the phenomenon occurs.
4. Increasing “fitness” (average reproductive output of a gene or genotype) is not the key to adaptive evolution. In the same link as the “no-cats-from-dogs” statement, Shapiro basically dismisses the importance of natural selection:
Was Darwin simply mistaken about the gradual nature of hereditary variation? Such ignorance would be unavoidable before we knew about Mendelian genetics and DNA. Or was there a deeper flaw in the theory that he (and Alfred Russell Wallace) propounded? The answer may well be that it was a basic mistake to think that optimizing fitness is the source of biological diversity.
What he’s saying, of course, is that natural selection (“optimizing fitness”) had nothing to do with the diversity of life.
This is all deeply misguided, and I suspect that Shapiro simply doesn’t understand natural selection. He certainly hasn’t proposed an alternative theory that explains all the adaptations we see in nature, and merely ascribes them in some nebulous way to the self-tinkering of the genome. But none of the mechanisms he adduces (save horizontal gene transfer, which is really just a big mutation, since that sort of transfer need not automatically be adaptive) can replace natural selection.
In his latest columns at HuffPo, (Part 1 and Part 2), Shapiro makes the same mistake, assuming that some features of the genome—the vertebrate immune system in this case—shows that natural selection is ineffective in molding adaptive traits of organisms, and that the innate nature of the genome has really replaced the conventional view of adaptive evolution. He assumes here that something as complex and “directed” as the vertebrate immune system could not have evolved by random mutation. What Shapiro fails to realize here is that these “innate features of the genome” that produce the appearance of “directed change” are themselves molded by a combination of random mutation and natural selection, creating a genome that operates in an adaptive way.
Let’s take a look at the way the immune system adaptively responds to challenges: in this case the presence of foreign proteins (“antigens”), such as those found on the surface of parasites or microbes, that must be neutralized. At the risk of doing short shrift to what is a breathtakingly impressive (and evolved) mechanism for fighting off disease and parasites, the vertebrate immune system operates like this:
- A foreign substance (or protein) invades the body. These substances are called antigens. (The coat proteins of bacteria and viruses are examples of such antigens). They pose a danger to the host.
- The body recognizes the antigen as foreign and swings into action to neutralize it. It does so by allowing a group of specialized cells (“B cells”) to generate a huge variety of different molecules, called antibodies, to neutralize the antigens. Antibodies are glycoproteins (large proteins with an attached carbohydrate) whose sequence is determined by the host’s DNA.
- The amazing thing is how the body uses a small number of genes in the B cells to generate a huge variety of protective antibodies, for we never know what foreign molecule is going to enter our bodies. (Antibodies are specific to antigens: not just any antibody molecule will neutralize an antigen.) What happens is that there are two processes, called somatic hypermutation and VJ recombination, that take the DNA sequence of the antibody-producing genes and mutate it, either creating “errors” in the DNA sequence or swapping bits within and among genes by physical recombination. This generates a large number of variable antibody proteins. Most of these won’t be useful for neutralizing the invading antigens, but some will.
- The “useful” antibodies, produced by random mutation and recombination, bind to the invading antigens; this alerts yet another cell type to migrate to the bound antigen-antibody complex and destroy the invading cell/protein.
- In another amazing part of this process, those cells that have undergone the “right” mutations and recombination (i.e., those whose DNA has changed in a way that produces antibodies useful to the invader) are induced by feedback from the antigen-antibody complexes to differentially proliferate in the body. Cells having the “right” mutations are stimulated to divide more often. This allows them to produce more of the useful antibody. Thus the population of the “good” immune cells is enriched, through differential replication, when they have a useful effect. I haven’t followed the literature on this, but I’m not sure we know exactly how the production of a successful antibody feeds back to the cells to induce them to divide more often.
- The reader will have recognized, as does Shapiro, that this differential proliferation of cells with the “right” mutations is a form of natural selection: the replicators (cells) that produce the most adaptive protein are those that proliferate, enriching the body in those cells and helping fight off the invader. Moreover, those cells remain in our body, which is why we can fight off infections more easily when we’re reinfected with the same antigen a second time. This is in fact the way vaccinations work: by giving us a mild form of the invader (or disease), which allows our bodies to produce more of the cells that can later swing into action when a real invader comes on the scene. Vaccinations involve “priming” the immune system with a small amount of antigen (i.e, the cowpox virus in the case of our first vaccination—Jenner’s vaccination for smallpox), allowing us to be prepared with extra antibody when the genuine infection comes. This use of our body’s evolved defenses by medical research is one of the great triumphs of the human mind.
This is all well and good, and Shapiro describes the process clearly in his first post (the second is way, way too complicated for the general reader). But his mistake comes when he assumes that the “adaptive” process of the immune response itself did not arise by the normal process of natural selection (random mutation followed by selection among genetic variants), but is somehow an inherent adaptive feature of the genome that renders “normal” natural selection unimportant.
Note that there are two levels of selection going on here. The first is the immune response itself: mutations (random ones, Dr. Shapiro!) occur in B cells, and those cells with the “right” mutations are caused to differentially proliferate. This is “somatic” natural selection; that is, selection within one generation and within one body to produce an adaptive result. But the appearance of the system itself requires a second form of selection: “regular” natural selection among ancestral vertebrates. In this process, those individuals having rudimentary immune systems better able to generate adaptive variants leave more offspring than other individuals whose immune systems can’t produce the variety of molecules needed to fight off a diversity of invaders.
The key point is that there is no evidence that the evolution of the immune system in this way (by differential reproduction of individuals instead of cells) involved anything other than natural selection among individuals having randomly produced mutant variants of an ancestral immune system. In other words, the whole system of hypermutability within B cells, and the feedback mechanism that allows cells with the “right” mutations to divide more often, arose by selection among individuals having slightly different forms of the immune system. In the end, the whole system appears “designed” to fight off invaders—even “designed” to produce the right mutations for the right invader, even though that’s not the way it works—but yet that whole system evolved by the usual step-by-step Darwinian process within ancestors. What is the alternative, Dr. Shapiro? Did God bequeath us a fully-functioning immune system de novo? I think not.
Evidence that such a system can indeed evolve comes from parallels within bacteria. Bacteria that are invaders themselves have their own process of producing antigens that avoid detection by the invader’s antibodies. In this process, bacteria have “hypermutable” genes called contingency loci that can produce new coat proteins that aren’t recognized by the somatically evolving vertebrate immune system. (This is the result of an “arms race” between the bacteria and the host’s immune system.) The evolution of contingency loci comes from pretty much the same mutational process responsibile for the evolution of hypermutability of B cells—both are evolved phenomena. Note especially that in both the contingency loci of bacteria and the hypermutability loci of vertebrates, the mutations that occur are random: variants are produced regardless of whether they’d help the beleaguered vertebrate trying to destroy the antigens or the besieged bacterium trying to avoid antibodies. But in both cases the cells producing “adaptive” responses (in the case of bacteria, the coat proteins that evade the host’s antibodies) undergo differential reproduction.
That is natural selection, pure and simple: randomly produced genetic variants followed by sorting out of those variants based on their ability to render their carriers more likely to survive. But somehow Shapiro manages to convey to the public that something else is going on here—some non-Darwinian process that can’t be explained by natural selection. And that’s bunk.
In fact, Shapiro sees my natural-selection scenario above, which is accepted by virtually every biologist who works on the immune system or on contingency loci, as a “philosophical” difference between him and me:
“Mutational change occurs by accident, or as a byproduct of something else (like a gene being accidentally duplicated, or the ingestion of DNA from another species), but those changes occur whether or not they’d be “good” (i.e., increase the reproductive output of) individuals in the species that has mutated.” We will return to the importance of this philosophical difference later in the blog.
It’s not a philosophical difference, it’s a scientific difference, and all the facts are on my side. There’s no evidence that the immune system resulted from anything other than garden-variety natural selection—the same sort of selection that helps bacteria evade the defenses of their hosts. Yet Shapiro claims that the immune system is “excluded from the prevailing philosophy of genetic change” (i.e., natural selection):
Three remarkable things about somatic hypermutation and CSR are explicitly excluded from the prevailing philosophy of genetic change. First, they are adaptive and purposeful genome changes. Second, they are functionally targeted. Third, for CSR, targeting involves intercellular signals that depend on how other cells in the immune system perceive a particular infection.
Yes, the changes are adaptive, and if you want to be metaphorically teleological about it, yes, they’re “purposeful”: the “purpose” of the changes is to fight off invaders. (I never use the word “purpose” when I lecture to students, as it implies a designer.) But equally “purposeful” are the adaptations of our ability to tan, or the ability of our cats to grow longer fur in winter.
The functional targeting and intercellular signals are simply part of the system that evolved. There’s nothing about this system that can’t be explained by natural selection, just as the “hypermutability” of bacterial contingency loci fall within the ambit of natural selection. If selection is strong enough, you can indeed evolve a system in which most of the products (e.g., the mutant B cells) produce useless proteins, for that hypermutable system still pays off often enough (i.e., generates the right molecule) that on the whole it increases the organism’s fitness. If you have a gene that produces five bad mutations for every good one, that gene can still be positively selected if the advantage of the rarer good mutation is strong enough to outweigh the detriments of the more numerous bad ones.
In the end, Shapiro pulls a scientific boner by saying that if the immune system evolved in such a complicated way, other cells should have too. But this is a red herring:
If immune cells can do all the above, is there any scientific reason we would assume that other cells cannot do the same? Coupling DNA restructuring to transcription is of major significance. All cells can target transcription to functionally relevant sites in the genome. Given that the immune system is how evolution evolved rapid protein evolution, should we not look to it for clues about basic evolutionary processes?
The burden of explaining what other cells lack that lymphocytes possess lays [sic] with those who wish to adopt the position that the immune response is unique and does not reflect a more general capacity to target genome change. Evolution has obviously refined antibody-producing cells for their immune system functions. But do immune cells have unique capacities for natural genetic engineering missing in other cells? If the answer is no, as I believe, then we need to incorporate adaptive genome restructuring into our most fundamental thinking about biology.
Well, we’ve seen that the immune system is not unique in its ability to “target” genome change: contingency loci in bacteria operate in the same way. But just because a complicated feedback system has evolved in one group does not mean that exactly the same mechanism will evolve in all cells, or in all groups of organisms. Indeed, we know that’s not the case. Insects’ resistance to organophosphate insecticides, for instance, has evolved by the simple modification of an enzyme that renders the insecticides ineffective. Surprisingly, it’s often the very same single mutation in different groups of insects. There’s no hypermutability here, no complex organic feedback, no “adaptive genome restructuring” or “natural genetic engineering.” It’s just a random mutation that happens to render the insect impervious to pesticides.
I am not sure why Shapiro is so obtuse about this, for he’s been trumpeting these same misconceptions for decades. My own theory is that the man simply doesn’t understand the kind of population thinking in which “natural genetic engineering” can result from garden-variety natural selection. (I often find that molecular biologists fail to grasp natural selection, even though it seems conceptually simple.) At any rate, Shapiro’s claims in HuffPo are damaging to the public understanding of science, for they make people think, unjustifiably, that there’s something very wrong with modern evolutionary theory. Well, his arguments aren’t convincing to biologists, although they could perhaps snow the layperson with complex terminology, just as Michael Behe snows the public with the idea of “irreducible complexity.”
The onus is on Shapiro to show exactly how the systems of “adaptive genome restructuring” he so admires require us to abandon our notion of adaptation via natural selection. He hasn’t convinced me, nor, as I’ll show in a subsequent post, one reviewer of his new book on this topic.