One often hears the suggestion that the neo-Darwinian view of evolution is on the skids, and that that view will be completely changed—if not overturned—by new biological ideas like modularity, genetic assimilation, evolvability, and epigenetics. Epigenetics in particular (I’ll define it in a moment) has been especially touted as a concept that will revolutionize evolutionary biology.
Call me an old fogey, but I think the idea of epigenetics as a Darwin-destroyer is completely bogus. Although certain discoveries in that area are interesting, and have certainly expanded our notion about how genes work, there is not the slightest evidence that the findings of epigenetics will dispel the main ideas of neo-Darwinism, which include the ideas of evolutionary change via natural selection and genetic drift, the randomness of mutations, the ideas of speciation and common descent, and the gene-centered view of evolution. I’ve explained my views on epigenetics as a revolution in several previous posts, for example here, here, here, and here, but, like the Lernean Hydra, each time you cut off a head of the epigenetic beast, it grows another one.
The latest head appeared in Friday’s Guardian, in a book review written by Peter Forbes; the book is The Epigenetics Revolution by Nessa Carey, and Forbes sees the book as tremendously important, implying that is part of a scientific revolution and explicitly saying that it’s a book that would “make Darwin swoon.” I haven’t read the book, and although it might make Darwin swoon if the old git were to be resurrected, the discoveries of genetics and the mechanism of inheritance itself would make him swoon far more readily. And I know scientific revolutions; scientific revolutions are friends of mine; and believe me, epigenetics is no scientific revolution.
So what is epigenetics? The term is actually used in two different ways. When I was younger, it simply meant “developmental genetics,” that is, the study of the way the DNA code of genes is translated into the bodies and physiologies of organisms. That, of course, was a tremendously important and exciting area, and still is. It involves understanding how genes are turned on and off in different tissues and cells, how different genes interact with each other, and how the products of a one-dimensional sequence of information can build a three-dimensional body. This study has segued into the new field of “evo devo,” which tries to understand the evolutionary basis of developmental genetics. “Evo devo” itself has, of course, led to its own important discoveries, like the presence and conservation of homeobox genes, the use of the same genes over and over again in forming similar but non-homologous traits (e.g., PAX6 in the formation of fly eyes and vertebrate eyes), and the linear arrangement of genes in some organisms (e.g., Drosophila) that correspond to the linear arrangement of body parts they affect.
So developmental genetics, and evo devo, are fascinating areas that produce a stream of surprising discoveries. But they’ve done nothing to alter the going paradigm of neo-Darwinian evolution. It is telling that, for example, Sean Carroll, a famous practitioner of “evo devo” and a popular writer, is a firm adherent to neo-Darwinism. What we learn from these areas is precisely how evolution has acted to sculpt bodies, but it still does so using randomly-generated genetic variation and good old natural selection (and yes, Larry Moran, genetic drift also plays a role). Gene regulation itself is a phenomenon molded by natural selection, and how genes are turned on and off is itself a phenomenon residing in the genes: in the genes that make the DNA or proteins that regulate other genes, and in the many ways that genes evolve (through, for example, the evolution of regulatory regions), to respond to internal “environmental” influences.
The second meaning of “epigenetics” is more recent, and involves actual changes in the DNA itself that are not based on mutational changes in nucleotides, but in environmental modifications of nucleotides—things like methylation of nucleotide bases or changes in DNA-associated proteins like histones—that can temporarily modify genes and affect their actions. I say “temporarily,” because such environmental modification of DNA, while it can be adaptive, is not usually passed on from one generation to the next. For example, we get our genes in pairs—one from mom and one from dad—but they can be differentially “marked” (the technical term is “imprinted”) during the formation of sperm and eggs, and so the copy from dad can act differently from the copy coming from mom. This imprinting is probably due to natural selection: scientists like David Haig have argued that the different and conflicting “interests” of paternal versus maternal genes has, through natural selection, molded the way they are imprinted, allowing them to act in different ways in the embryo. But an “imprinted” gene is reset each generation: the imprinting disappears and has to re-form depending on which sex the gene is in.
As I have argued before, however, imprinting of genes, although a novel and recently-discovered phenomenon, is not a “revolution” in how we view evolution: it is an embellishment that doesn’t overturn the main ideas of neo-Darwinism. And many of the phenomena subsumed by this modern notion of “epigenetics” still evolved by natural selection. Imprinting, after all, is based on changes in DNA that somehow render paternal DNA more (or less) susceptible to modification than maternal DNA. Imprinting has evolved by changes in DNA, even though the modifications of DNA it causes are environmental.
In his review of Carey’s book, Forbes, a science writer, concentrates on the second, “revolutionary” sense of epigenetics:
Genes don’t just issue instructions: they respond to messages coming from other genes, from hormones and from nutritional cues and learning. Much epigenetics revolves around nutrition. If we drink a lot of alcohol an enzyme that metabolises it becomes more active – “upregulated” in the jargon. And similar mechanisms apply to much of our behaviour. The methods by which genes makes these responses often involve very small chemical modifications (usually the addition of a tiny methyl group to one base of DNA). It is almost certain that memory – a classic nurture problem: we learn something and it becomes biologically encoded – involves epigenetics. Once made, epigenetic changes can be very long lasting, which is how our long-term memory is possible.
Why is this “revolutionary?” Because some of the inherited changes of genes appear to be “Lamarckian,” that is, they aren’t really changes in DNA sequence itself, but environmental modifications of DNA that can be passed from one generation to the next. And if such “nongenetic,” environmentally-acquired inheritance were common, that would be a revolution in the way we think about evolution.
So what’s the evidence for this “revolutionary” notion? Forbes simply offers up the same tired old anecdotes I’ve addressed before:
So far, this is instructive and highly promising for medical research, but epigenetics finally reaches that “everything you’ve been told is wrong” moment when it claims that some epigenetic changes are so long-lasting they cover several generations: they can be inherited. This flouts one of biology’s most cherished dogmas – taught to all students – namely that changes acquired during life cannot be passed on – the heresy of Lamarckism.
But the evidence that this can occur in some cases appears to be growing. There are lab experiments with mice and rats in which epigenetic effects on coat colour and obesity can be inherited. More suggestive evidence comes from a vast, unwitting and cruel experiment played out in the second world war. In 1944, during the last months of the war, a Nazi blockade followed by an exceedingly harsh winter led to mass starvation in Holland. This had a huge effect on babies born at the time, and the effects of poor nutrition on the foetus seem to have persisted through subsequent generations.
Well, I won’t flog poor Mr. Forbes with the fact that these are only a few trivial examples of the phenomenon, examples that don’t appear to have any evolutionary importance. Nor will I flog him with the fact that when we can dissect the genetic basis of real adaptations in real organisms, they invariably turn out to rest on changes in DNA sequence, not in environmental and non-DNA-based modifications of nucleotides. Here’s what I said in an earlier post about Oliver Burkeman’s claims that epigenetics has profound implications for evolution (like Forbes, Burkeman is a science journalist):
All I can say to this is: “Profound implications my tuchus!” There are a handful of examples showing that environmentally-induced changes can be passed from one generation to the next. In nearly all of these examples, the changes disappear after one or two generations, so they couldn’t effect permanent evolutionary change. The proponents of epigenesis as an important factor in evolution, like Eva Jablonka and Marion Lamb, always wind up talking about the same tired old examples, like cases of coat color change in mice and flower pattern in toadflax. I am not aware of a single case in which an adaptive change in an organism—or any change that has been fixed in a species—rests on inheritance that is not based on changes in the DNA. (For a refutation of the pro-epigenesis arguments that Jablonka and Lamb make in their 2005 book, see Haig .)
Moreover, some examples of “nongenetic” inheritance that do have adaptive significance, such as differential methylation of paternal versus maternal chromosomes, ultimately rest on changes in DNA that promote that differential methylation. And this “inheritance” lasts only one generation, for the methylation profile is reset in each sex every generation.
In contrast to the very few cases of one- or two-generation inheritance that cause nonadaptive changes in the phenotype stands the very, very large number of studies in which inherited changes within and among species map to the DNA. These include every case of evolutionary response to artificial or human-generated selection, adaptive changes within species (e.g., spiny-ness in sticklebacks), and differences among species in both morphology (e.g., the color differences in fruit flies I study) and reproductive barriers (the many mapping studies of “hybrid sterility” and “hybrid inviability” genes). Burkeman, of course, doesn’t mention these cases: it would ruin his nice story.
If we look just at studies of the inheritance of organismal changes that have evolved over time (and many of these would have detected profound epigenetic effects), the score would be something like this: DNA 757, Epigenesis 0. (I’m just making these numbers up, of course, to make a point.) If we look at all “inherited changes”, regardless of their evolutionary importance, we would have a handful of epigenetic changes versus literally thousands of DNA-based changes. So how can Burkeman say that epigenesis will profoundly revise our view of evolution?
So, Mr. Forbes, our “cherished dogma” of non-Lamarckian inheritance still holds strong, and you’ve done your readers a disservice by implying otherwise. Lamarckism is not a “heresy,” but simply a hypothesis that hasn’t held up, despite legions of evolution-revolutionaries who argue that it flushes neo-Darwinism down the toilet. If “epigenetics” in the second sense is so important in evolution, let us have a list of, say, a hundred adaptations of organisms that evolved in this Larmackian way as opposed to the old, boring, neo-Darwinian way involving inherited changes in DNA sequence.
Forbes can’t produce such a list, because there’s not one. In fact, I can’t think of a single entry for that list.
Science journalists—meh. They’re always trying to argue that Darwin was wrong and that evolution is about to undergo a Kuhnian revolutionary paradigm. But what they really want is readership, and you don’t get readers by writing that the conventional wisdom happens to be correct.