Epigenetics kerfuffle

Last week I highlighted a piece on epigenetics by Florian Maderspacher, an editor at Current Biology.  (For this post I’ll define “epigenetics” as “inheritance that transcends generations but is not based on changes in DNA sequence”.) Florian decried all the current hype about epigenetics—that is, the idea that this phenomenon will revolutionize our view of evolution—and mentioned its connection to Trofim Lysenko, the Soviet agronomist who wrought immense havoc on Russian science and agriculture with his misguided notion that environmental modifications of crops could become permanently fixed in their genes.

I agree with much of what Maderspacher said (I’ve posted on this before): the “evolutionary” notion of epigenetics is an overblown construct with little empirical backing, and seems largely to further the career aspirations of its advocates.

Now the two biggest drivers of the “evolutionary epigenesis” bandwagon, Eva Jablonka and Marion Lamb, have replied in Current Biology, and Maderspacher responds (scroll down when you get to the page).  Jablonka and Lamb object to the Lysenko analogy as tarring their new paradigm with the ghosts of pseudoscience past.  Maderspacher responds:

Of course, tastes are allowed to differ, but I firmly believe that in scientific discourse, which is not categorically different from other forms of discourse, there is some room for stylistic freedom, all the more in an opinion piece such as mine. If you compare my six-column piece to the amount of paper that has been filled with sensationalism about epigenetics, it seems legitimate that the ‘small dog’ barks a little louder, as it were. (that in the interim both science and nature neuroscience have run special issues on epigenetics is a case in point)

I chose Lysenko as an admittedly extreme figurehead, because he exemplifies what can happen if political interpretation gets in the way of scientific rigour. In particular I was interested in the interpretation the public brings towards a rather esoteric field of molecular biology and genetics. Why are people so interested in epigenetics? Because they like to see it as a liberation (”victory over the genes”), and this is exactly where Lysenko was coming from. I wanted to expose this distorted reasoning. To imply that I equate the study of environmental influences on heredity with Lysenkoism is, frankly, untenable.

But of course this part of the kerfuffle is about framing and tone. The more important thing is the science behind the epigenetics hype.  And I maintain, as I have for a while, that there is simply no good data supporting the idea of non-genetically based and transgenerational inheritance as an important factor in evolution.  I’ve read nearly all of Jablonka and Lamb’s papers, and they keep recycling the same tired old examples (mouse fat and toadflax) to show how drastically epigenetic inheritance can alter our view of neo-Darwinian evolution.  They have not said anything new for a long time.

Their arguments are unconvincing for a number of reasons. Epigenetic inheritance, like methylated bits of DNA, histone modifications, and the like, constitute temporary “inheritance” that may transcend one or two generations but don’t have the permanance to effect evolutionary change.  (Methylated DNA, for instance, is demethylated and reset in every generation.)  Further, much epigenetic change, like methylation of DNA, is really coded for in the DNA, so what we have is simply a normal alteration of the phenotype (in this case the “phenotype” is DNA) by garden variety nucleotide mutations in the DNA.  There’s nothing new here—certainly no new paradigm.  And when you map adaptive evolutionary change, and see where it resides in the genome, you invariably find that it rests on changes in DNA sequence, either structural-gene mutations or nucleotide changes in miRNAs or regulatory regions.  I know of not a single good case where any evolutionary change was caused by non-DNA-based inheritance.

Maderspacher hits the nail squarely:

This leads to the second line of criticism, my allegedly distorted portrayal of epigenetics as a scientific discipline. I happily admit that I am not an epigenetics expert, by any stretch of the imagination. I would even argue that it can sometimes be helpful to have a view from afar, from someone who has no personal stakes in the matter. And I certainly agree that paradigms (and terms) in biology shift all the time. But they shift when new empirical data call the pervious paradigm into question. In the case of DNA and histone modification, I cannot see how they call for a change in our view of genetics. As for epigenetic inheritance, I am aware that there are numerous examples and mention some in the piece. This is interesting biology and deserves attention. But does it call for a paradigm shift? Only time will tell, I guess.

I’m talking in Medellin, Colombia next week about all of the new “challenges” to neo-Darwinism, and epigenesis is among them.  Relevant to the increasing cacophony from the “Darwin-was-wrongers,” I have two quotes.  The first is from evolutionary biologist Doug Futuyma:

One has to have a certain degree of reservation about claims that are made on the basis of one or two examples that are going to be a major challenge or a new expansion.  Otherwise you’re talking about jumping on one  bandwagon after another.

The second is from my Ph.D. advisor Dick Lewontin:

. . . scientists are always looking to find some theory or idea that they can push as something that nobody else  ever thought of because that’s the way they get their prestige. . . .they have an idea which will overturn our whole view of evolution because otherwise they’re just workers in the factory, so to speak. And the factory was designed by Charles Darwin.

Florian is much more polite than I am, but as I age I become less reluctant to broach the charge of careerism.  Being ambitious is not inherently bad: all scientists secretly wish to get a big name and be lionized among their peers. A desire for renown—to be first with the goods—has motivated much of the best science we have.  But there’s a good way and a bad way to get famous.  The good way is to produce solid, interesting data and avoid overhyping them.  The bad way is what Lewontin was talking about above.


  1. Posted November 4, 2010 at 7:06 am | Permalink

    And Jablonka & Lamb have indeed been successful at “selling” their “product” to the public (which as I understand it makes research grants easier to get). Just the other day, my wife said, “Hey, this is interesting… an article about how obese mice can pass that on to their offspring…” Luckily, being a regular reader of WEIT, I said, “Hey, I know those guys! Yeah, nobody takes them seriously…” heh..

  2. NewEnglandBob
    Posted November 4, 2010 at 7:08 am | Permalink

    …because otherwise they’re just workers in the factory, so to speak. And the factory was designed by Charles Darwin.

    A brilliant comment.

  3. Tim Martin
    Posted November 4, 2010 at 7:42 am | Permalink

    I just have a question about the idea that epigenetic inheritance has to be temporary. Here’s something I read from a short article on epigenetics:
    “Champagne’s research shows that, via methylation, high versus low levels of licking and grooming influence the expression of a gene that governs estrogen receptors in a key brain area responsible for maternal behavior. In this way, the amount of nurturance received by a mouse pup influences that pup’s own maternal behavior later on…”

    The way one mouse cares for its pups affects their brains in such a way that it causes them to treat their pups the same way. It seems to me that this sort of thing could, under the right circumstances, go on indefinitely. Is that not the case? Would the effect tend to diminish over several generations?

    • Tulse
      Posted November 4, 2010 at 8:10 am | Permalink

      The way one mouse cares for its pups affects their brains in such a way that it causes them to treat their pups the same way.

      That may be true, but it’s important not to get to carried away. After all, children who grow up in English-speaking households tend to speak to their own children in English. (And, for that matter, if the issue is nurturance, many abused children grow up to be child abusers themselves, but it’s probably not due to methylation.)

      “Learning” is a form of epigenetics (depending on how one defines “inheritance”), but it doesn’t rely on the typically-proposed molecular epigenetic mechanisms.

    • Posted November 4, 2010 at 8:36 am | Permalink

      The other point is that whether you are talking about “behavioral epigenetics” as we might call it, or molecular epigenetics, either way the inheritance is far less “sticky”. If you’ve got a particular allele, by gum you’ve got that allele, and if you have offspring, each one has a (roughly) 50% chance of getting that allele, regardless of anything you do or anything that happens to you. These sorts of epigenetic mechanisms, on the other hand, even if they have a sort of positive feedback effect that tends to make them persist for a number of generations, they can be disrupted trivially.

      Remember, we are talking about timescales of hundreds of thousands of years here. “Grandma mouse groomed her pups nicely, so mama mouse did too, and now so does baby mouse” sounds nice, but it’s not even a blip on an evolutionary timescale.

      • Tim Martin
        Posted November 4, 2010 at 1:01 pm | Permalink

        Gotcha. Thanks, James and Tulse.

  4. Posted November 4, 2010 at 8:17 am | Permalink

    In a recent review by Jablonka and Lamb, they included a table with well over 100 examples of what they call “transgenerational epigenetic inheritance in prokaryotes and eukaryotes.” ( http://www.ncbi.nlm.nih.gov/pubmed/19606595 ) Their examples range from bacteria to animals to plants. You say “they keep recycling the same tired old examples (mouse fat and toadflax.” Could you clarify what you mean by “examples”? Are all the 100+ cases listed in their recent review the examples you’re referring to? Or are some of the cases they’re talking about irrelevant to the questions of whether epigenetics is a significant challenge to neo-Darwinism?

    I just want to understand the terms of the debate better.

    • whyevolutionistrue
      Posted November 4, 2010 at 8:27 am | Permalink

      Maternal effects are like these, too, and most examples in that paper (save for those in microorganisms, which resemble “epigenetic” inheritance among cells within a metazoan) last only a handful of generations, and usually disappear during meiosis. Therefore they’re irrelevant to evolutionary change. Further, many of them apparently rest on coding changes in the DNA (methylation, chromatic marking, etc.) I never denied that environmental changes could sometimes be passed between generations, just that these are a. semipermanent forms of inheritance, b. independent of changes in DNA sequence, and c.known to be important in evolution.

      The mouse-fat stuff (and maybe the toadflax stuff) lasts more than a generation, so that’s why they tout these as potentially important for evolution. (Read their papers; you’ll soon get tired of mouse-fat and toadflax!). But even these effects disappear eventually. And they’re not adaptations, just genetic anomalies. Granted, those could be important in adaptation, theoretically, but where are all the examples? Where is ONE example?

      Environmentally induced changes that last only one or a couple of generations cannot be a basis of evolutionary change. Giving many examples of such transgenerational changes doesn’t convince me of anything.

      I can steer you to some papers on this topic by David Haig if you wish.

  5. MoonShark
    Posted November 4, 2010 at 8:32 am | Permalink

    I think epigenetics is important and interesting because of its possible influence on human behavior and heath (in a similar manner to how gut flora make us each unique and interesting). After all, genetically identical twins aren’t exact doppelgangers, sharing the same thoughts and personality characteristics. It’s worthwhile to understand the “environment” part of the equation in addition to the genes.

    But yeah, I’m somewhat skeptical of the idea that methylation and histone modification are important to evolution itself. That might well be a stretch. Instead it’s an interesting *supplement* to evolution, which may help us explain individual differences among otherwise genetically similar or even identical organisms.

    • ritebrother
      Posted November 4, 2010 at 8:47 am | Permalink

      A good bit is known about the underlying differences in epigenetic marks (histone mods & DNA me’n) between twins, and how the disparity increases with time and woth increasing lifestyle disparity. See the work of Manel Esteller.

      • MoonShark
        Posted November 4, 2010 at 8:56 am | Permalink

        Sounds interesting; I’ll have to check that out. Thanks!

  6. theshortearedowl
    Posted November 4, 2010 at 8:44 am | Permalink

    Plus, surely the epigenetic changes themselves are also genetically coded? As in, there will be variation in degree and patterns of epigenetic DNA methylation between individuals in response to the same environmental variables, and at least some of that variation can be explained by genetic variation?

    • ritebrother
      Posted November 5, 2010 at 11:13 am | Permalink

      Indeed. Epigentic marks are effected by protein gene products, so alleles of these enzymes can affect DNA and histone modifications, and thus epigenetic effects. Many of these genes were identified in mutant screens in yeast and Drosophila based on very significant phenotypes as modifiers of other effects.

  7. Posted November 4, 2010 at 9:01 am | Permalink

    I agree with most of this article about overhype, but this is poorly reasoned:

    “…last only a handful of generations. Therefore they’re irrelevant to evolutionary change.”

    They last only a few generations after you remove the environmental stimulus that induced the epigenetic change. In the continuing presence of an environmental factor (not hard to imagine), the epigenetic difference would persist as well. It is still not permanent, but because it affects phenotype has the potential to alter selection dynamics, driving “real” evolution through changes in allele frequencies.

    I agree 100% that there is no strong evidence that this type of process has played an “important” role in evolution of the tiny percentage of organisms who happen to be multicellular eukaryotes with chromatin modifying systems, but it’s not an incoherent idea.

    And James Sweet…the obese mice altering the metabolism of their offspring and other maternal/paternal effects are not particularly controversial phenomena (and I don’t think Jablonka & Lamb have anything to do with it, so not sure who “those guys” are). Whether it’s relevant to species evolution or not is beside the point–it’s a form of inheritance and has clear public health implications.

    • ritebrother
      Posted November 4, 2010 at 9:48 am | Permalink

      It’s largely the work of Randy Jirtle, Robert Waterland, David Martin and Emma Whitelaw.

  8. Dominic
    Posted November 4, 2010 at 9:51 am | Permalink

    On similar lines, & related to the did you see the New Scientist (yes, I still et it) article a few weeks ago by Keith Bennett of professor of late-Quaternary environmental change at Queen’s University Belfast?

    I warn you that it comes down on the side of Fodor/Piattelli-Palmarini –
    ‘In the last analysis, evolution can be likened to the description of human history as “just one damn thing after another”, exactly as Fodor and Piattelli-Palmarini have argued.’


  9. Kevin
    Posted November 4, 2010 at 9:59 am | Permalink

    What I don’t understand is why epigenetics is suddenly so fashionable (although not fashionable enough that my spell checker doesn’t give the word the “squiggle”).

    I just did a PubMed search on “epigenetics” and got 39,412 hits, dating back to 1981.

    Surely by now and after 39,000+ papers on the issue over 30 years, we’d know if this phenomenon was going to overturn our understanding of evolution.

    That it hasn’t doesn’t mean it won’t. But my money’s on the other side.

    • ritebrother
      Posted November 4, 2010 at 10:09 am | Permalink

      It’s fashionable because molecular biologists (full disclosre: I am one) co-opted the term to specifically refer to the chemical modifications of histones and DNA per se. It was originally meant to encompass ALL mechanisms that affect gene expression, and thus phenotype. This would originally include transcription factors, which is why Mark Ptashne is always railing against its current definition (I have some references if interested). Others, such as Adrian Bird (and most of the chromatin & transcription field), have conceeded to the redefinition.

      • Kevin
        Posted November 4, 2010 at 10:18 am | Permalink

        Ah. So a semantics fight along with a scientific argument about the importance of a phenomenon/epiphenomenon?

        No wonder I couldn’t make sense of the issue. I was researching epigenetics and you were discussing epigenetics.

        And with that, I will back slowly out of the room.

        • Kevin
          Posted November 4, 2010 at 10:21 am | Permalink

          Except one more question. What do we now call the regular plain vanilla epigenetics? Epiepigenetics? Mepigenetics? Shmepigenetics?

          If you grab a useful term and turn it into something specific, you can’t just abandon the rest alongside the road, broke and hungry.

          • Posted November 4, 2010 at 11:23 am | Permalink

            Part of the problem, in my estimation, is that some of the ‘new’ epigenetics folks want to have their cake and eat it too. They imply that they mean chemical modification, but then use the word in such a broad sense to basically mean ‘gene expression.’ It’s hard to grapple with something when the definition is slippery! Depending on the context, or even the author, it epigenetic inheritence could be transgenerational, within a generation, within a portion of a life history…

            And that’s leaving aside the whole mess left by people who do methylation screening irresposnibly.

            • ritebrother
              Posted November 4, 2010 at 11:40 am | Permalink

              This is my asessment of the situation as well.

          • ritebrother
            Posted November 4, 2010 at 12:33 pm | Permalink

            How about “metagenetics”? More seriously, I agree with you. The original Waddington concept of the unfolding of the developmental program was more fluid, and did not specify mechanisms. For example, gradients of signalling molecules that lead to spacially differential gene expression during ontogeny would also be “epigenetic”.

  10. slpage
    Posted November 4, 2010 at 10:03 am | Permalink

    Reminds me of the “directed mutation” flap years ago. Is Rosenberg still claiming that this phenomenon is real, I wonder….

  11. Josh
    Posted November 4, 2010 at 10:04 am | Permalink

    Is there any mathematical theory exploring the implications of epigenetic inheritance? I suspect that in a “random-environments” framework, epigenetic changes could cause the fixation of alleles that would otherwise hang around at intermediate frequency (c.f. Gillespie’s random environments models).

  12. Posted November 4, 2010 at 10:39 am | Permalink

    …they have an idea which will overturn our whole view of evolution because otherwise they’re just workers in the factory, so to speak. And the factory was designed by Charles Darwin.

    And then, of course, there’s Craig Venter, who wasn’t satisfied with Darwin’s factory and went and built his own. From scratch….



  13. Greg Peterson
    Posted November 4, 2010 at 11:41 am | Permalink

    I was really eager to read the book “Anarcy Evolution” by punk band Band Religion’s frontman Greg Graffin, who has a zoology PhD and teaches life sciences at UCLA. I’m a big fan of Bad Religion, and Graffin’s firm stance for reason and humanism and against authoritarianism and superstition. But as much as I liked parts of the book (especially the more autobiographical parts), his take on epigenetics struck me as sort of woo-ful. Or perhaps just not as sophisticated as I expected it to be.

  14. Zuropa
    Posted November 4, 2010 at 12:54 pm | Permalink

    I recently saw a TED talk by Heribert Watzke in which he argues that cooking drove our digestive physiology toward its present state. Is this an example of an epigenetic factor? This would also be a contributing explanation for the extreme cognitive biases which religion/culture exerts on the mind. Not only are are there benefits in terms of social cohesion, but our physiology actually depends upon information being passed down epigenetically, namely, the learned art of cooking.

  15. A.Way
    Posted November 4, 2010 at 6:59 pm | Permalink

    So – darwinian finches – was this evolution or epigenetics? Sure sound like epigenetics…

    • Luc
      Posted October 6, 2013 at 3:09 am | Permalink

      I once read that a recent research on darwin’s finches have shown that the responsible factor has been already isolated and even tests on chickens eggs succesfully conducted, but I can find the reference, so if some is aware of it;
      the word epigenetic wasn’t used but is perhaps a kind of academic auto-censure

      • Luc
        Posted October 6, 2013 at 3:15 am | Permalink

        I found it :
        Bmp4 and Morphological Variation of Beaks in Darwin’s Finches

        Arhat Abzhanov1,
        Meredith Protas1,
        B. Rosemary Grant2,
        Peter R. Grant2,
        Clifford J. Tabin1,*

        • A.Way
          Posted October 6, 2013 at 6:37 am | Permalink

          That article identified Bmp4 as associated with changed in beak morphology and change in expression of Bmp4 as maybe being the cause in change in beak shape. The change in expression can be an epigenetic effect. Examples are the Russian tame foxes and the Agouti mouse. You have changed in gene expression without a change in DNA base pairs. This article did not confirm a change in the genetic code.

          • Luc
            Posted October 9, 2013 at 2:38 am | Permalink

            for sure and this was my point, that evolution is mostly epigenetic, and that should perhaps explain the punctuated equilibrum stated by Gould

  16. Robert MacDonald
    Posted November 4, 2010 at 7:11 pm | Permalink

    I thought Canadians had kerfuffles, while everyone else had brawls and catfights.

  17. ambulocetacean
    Posted November 5, 2010 at 4:53 am | Permalink

    Do be careful in Colombia. It can be a bit rough in places.

  18. Posted November 5, 2010 at 6:29 am | Permalink

    Help me understand this better.It seems pretty clear that epigenetic changes do transmit to one, or two or three, who knows, generations, as was shown in a recent study on fattened rats.Why would this not have an impact on reproductive success, and therefore on inheritance?

    • ritebrother
      Posted November 5, 2010 at 7:02 am | Permalink

      Because those modifications (DNA methylation and histone modifications, which are the agents of epigenetic inheritance analogous to an allele) are not that stable, and are actively reset during gametogenesis and reestablished in a specific pattern in the developing embryo at each generation. The penetrance of these effects are also highly variable and contingent on environment (especially diet). Many of these effects require specific diets and levels of supplements (e.g. one-carbon pathway substituents like folate) to achieve significant results in mouse models.

      • Posted November 5, 2010 at 7:25 am | Permalink

        Interesting !But if a fat rat leads to a diabetic offspring rat, and that rat reproduces less, doesnt that mean that we have in effect a changed inheritance, even if its not written in DNA?

        • ritebrother
          Posted November 5, 2010 at 8:17 am | Permalink

          Yes, but there is no heritable agent that is selected, since the reduced fitness is not encoded by a discrete gene allele, despite the fact that phenotype affects fitness.

          • A.Way
            Posted November 5, 2010 at 10:20 am | Permalink

            Yes, but all other genetic aspects are having selection pressure if the organism’s fitness has changed. So epigenetics can affect selection.

            • ritebrother
              Posted November 5, 2010 at 10:48 am | Permalink

              Agreed. Any alteration in physiology will affect the fitness landscape against which gene alleles are selected. The question is whether the stability and penetrance of epigenetic effects are sufficient to really alter the fitness landscape for a significant population.

  19. Posted November 5, 2010 at 7:24 am | Permalink

    Two comments. First, Jerry, I appreciate that you always make clear you are using the word ‘epigenetics’ in the ahistorical sense favored by Jablonka, rather than in its primary meaning. 😉 Second, as regards career building, Pere Alberch once told me that John Maynard Smith (referring to punctuated equilibrium) said that Gould and Eldredge had read Kuhn’s Structure of Scientific Revolutions, and were following the instructions.

  20. Posted November 5, 2010 at 7:51 am | Permalink

    The use of the word decimate pained me. Unless the meteorite was only responsible for 10% of the dinosaurs extinction.

    Alas, I fear hubris will come my way, I make a poor pedant.

  21. anotherJoe
    Posted November 5, 2010 at 9:00 pm | Permalink

    Am I the only reader asking myself how flax is to toad as fat is to mouse? Consider the question rhetorical; I know how to google. There is also the hint of he missing hyphen.
    Usage can change or even reverse the meaning of words; the purist’s decimate is long gone. It is hard to even think of a correct use that doesn’t refer to a Roman military punishment, though I understand the USN submarine service was decimated in WWII; the Japanese and German services did much worse, I believe.
    I refuse to use the word myself, and give a little sigh when I see it, just as I refuse to speak of “quantum leaps” outside atoms.

2 Trackbacks/Pingbacks

  1. […] Epigenetics kerfuffle (whyevolutionistrue.wordpress.com) […]

  2. […] (More on that later.) It’s an argument that Jerry Coyne has refuted well on more than one occasion. In […]

%d bloggers like this: