Behe’s new paper

December 12, 2010 • 12:37 pm

The latest issue of The Quarterly Review of Biology has a paper by intelligent-design advocate Michael Behe, “Experimental evolution, loss-of-function mutations, and “the first rule of adaptive evolution.” It’s a review of several decades’ worth of experimental evolution in microbes (viruses and bacteria), with an eye toward revealing exactly what kinds of mutations have occurred in these studies.  He concludes that microbial evolution in the lab has been based largely on mutations that either 1) degrade or destroy functional elements like genes and promoter sequences, or 2) “modify” the function of pre-existing genetic elements so they do something slightly but not qualitatively different. What Behe does not observe is the evolution of “new functional elements” (see below): new genes, new coding sequences, new promoter regions, and the like.

When Behe produces a paper like this, it’s hard to resist imputing a motivation for the work.  After all, the man has a long history of promoting ID, and has written two books (which I’ve reviewed—negatively—here and here [click “God in the details” at the bottom of the page]) purporting to show that a celestial hand was necessary in evolution.  I’ve also caught Behe deliberately misquoting me in the service of his creationist views.  I believe—and think that time will prove me right —that his intention is to show that evolution cannot provide new structures or new “information” (e.g., genes), but can only either tinker with ones already present or degrade them.   Thus, to explain the evolution of truly new genetic information, one must invoke the intervention of an intelligent designer.

I think that while Behe’s summary of the results of these short-term lab experiments is generally accurate, one would be completely off the mark to extend his conclusions to evolution in general—that is, evolution as it has occurred in nature, be it in microbes or eukaryotes.

To make a long paper short, let me give the definitions Behe uses to reach his conclusion.

These are the things Behe doesn’t see arising in lab experiments:

Functional Coded elemenTs (FCTs): “An FCT is a discrete but not necessarily contiguous region of a gene that, by means of its nucleotide sequence, influences the production, processing, or biological activity of a particular nucleic acid or protein, or its specific binding to another molecule. Examples of FCTs are: promoters; enhancers; insulators; Shine-Dalgarno sequences; tRNA genes; miRNA genes; protein coding sequences; organellar targeting- or localization-signals; intron/extron splice sites; codons specifying the binding site of a protein for another molecule (such as its substrate, another protein, or a small allosteric regulator); codons specifying a processing site of a protein (such as a cleavage, myristoylation, or phosphorylation site); polyadenylation signals; and transcription and translation termination signals.”

In other other words, FCTs are new genes or new parts of genes, including those genetic changes that produce proteins with qualitatively new functions, not just a change in protein amount (e.g., changes in splicing or phosphorylation sites).

According to Behe, there are three types of adaptive genetic mutations that can be seen as differing in their effects on FCTs:

1.  “Loss-of-FCT” mutations. Those are the changes that have adaptive effects by destroying or degrading an FCT.  These include frame-shift mutations that render a protein inactive, or mutations that destroy a gene’s ability to bind to a transcription factor.

2.  “Gain-of-FCT” mutations.  These are the ones that Behe doesn’t see.  He defines them as mutations “that produce a specific, new, functional coded element while adapting an organism to its environment. The construction by mutation of a new promoter, intron/exon splice site, or protein processing site are gain-of-FCT mutations. Also included in this category is the divergence by mutation of the activity of a previously duplicated coded element.” In other words, mutations in this category produce new genes, parts of genes, or confer drastic new capabilities on genes by adding new splicing sites.  Also note that because almost no bacteria or viruses have introns in their cellular genes, it’s impossible to even see one class of this mutation in lab experiments on these groups.

3.  “Modification-of-function” mutations. These include every adaptive mutation that doesn’t fall in the above two categories, including point mutations that affect protein structure or quantitatively affect protein quantity, gene duplications that occur without sequence divergence,  rearrangements of gene order, etc.  He calls these “modification of function” rather than “modification of FCT” because the functional change doesn’t have to occur by changing an FCT itself.

Now these categories are not cut and dried.  For example, the “sickle-cell” mutation that, when present in one copy, protects carriers against malaria, is a point mutation in the beta hemoglobin molecule, changing a glutamic acid residue to a valine. You’d think that this would fall under class 3 (“point mutations”), but Behe considers it an adaptive gain of an FCT because the mutation causes the mutant hemoglobins to stick to each other in blood cells, somehow inhibiting the growth of the malaria parasite.  And because the point mutation is thereby said to specify a “new protein binding site”, Behe puts it into class 2 (gain of FCT).  Unfortunately, a lot of the single-gene mutations that Behe reviews from the experimental microbial-evolution literature work in unknown ways, so he could be missing similar cases that really fall into class 2.

Anyway, Behe reviews the last four decades of work on experimental evolution in bacteria and viruses (phage), and finds that nearly all the adaptive mutations in these studies fall into classes 1 and 3.  We see very few “gain of FCT” mutations.  Although this is not my field, the review seems pretty thorough to me, and the conclusions, as far as they apply to lab studies of adaptation in viruses and bacteria, seem sound.  From this Behe formulates what he calls “The First Rule of Adaptive Evolution:

Break or blunt any functional coded element whose loss would yield a net fitness gain.

What this means is that if adaptation can be gained by losing gene or enzyme activity, it’s more likely to occur by a loss-of-FCT mutation than the appearance of a new FCT itself with altered and reduced function.  That’s not really a “law” but a generalization from these lab experiments.

My overall conclusion: Behe has provided a useful survey of mutations that cause adaptation in short-term lab experiments on microbes (note that at least one of these—Rich Lenski’s study— extends over several decades).  But his conclusions may be misleading when you extend them to bacterial or viral evolution in nature, and are certainly misleading if you extend them to eukaryotes (organisms with complex cells), for several reasons:

1.  In virtually none of the experiments summarized by Behe was there the possibility of adapting the way that many bacteria and viruses actually adapt in nature: by the uptake of DNA from other microbes.  Lenski’s studies of E. coli, for instance, and Bull’s work on phage evolution, deliberately preclude the presence of other species that could serve as vectors of DNA, and thus of new FCTs.   This is not an idle objection, since we know that adaptation in natural populations of microbes often arises by incorporating new FCTs from other species.  Pathogenicity and antibiotic resistance in bacteria, for example, arise in this way.  Howard Ochman at Yale has done many studies on the acquisition of new bacterial functions by uptake of DNA from other species (and the source of the new DNA is often mysterious).

2.  In relatively short-term lab experiments there has simply not been enough time to observe the accumulation of complex FCTs, which take time to build or acquire from a rare horizontal transmission event.  Finding adaptation via point mutations or loss of function is much more likely.  Behe admits this much, but downplays it by saying this:

After all, one certainly would not expect new genes with complex new properties to arise on such short time-scales. Although it is true that new complex gain-of-FCT mutations are not expected to occur on short time-scales, the importance of experimental studies to our understanding of adaptation lies elsewhere. Leaving aside gain-of-FCT for the moment, the work reviewed here shows that organisms do indeed adapt quickly in the laboratory—by loss-of-FCT and modification-of-function mutations. If such adaptive mutations also arrive first in the wild, as they of course would be expected to, then those will also be the kinds of mutations that are first available to selection in nature. This is a significant addition to our understanding of adaptation.

and this:

A third objection could be that the time and population scales of even the most ambitious laboratory evolution experiments are dwarfed when compared to those of nature. It is certainly true that, over the long course of history, many critical gain-of-FCT events occurred. However, that does not lessen our understanding, based upon work by many laboratories over the course of decades, of how evolution works in the short term, or of how the incessant background of loss-of-FCT mutations may influence adaptation.

What he’s saying is this:  “Yes, gain of FCTs could, and likely is, more important in nature than seen in these short-term experiments.  But my conclusions are limited to these types of short-term lab studies.”  Well, good, but then let us not hear Behe’s ID colleagues tout these results as giving strong conclusions about microbial or eukyaryotic evolution in nature, particularly because the lab studies deliberately exclude sources of gain-of-FCT mutations that we know are important in nature.

3.  Finally, Behe does not mention—and I think he should have—the extensive and very strong evidence for adaptation via gain-of-FCT mutations in eukaryotes.  While that group may occasionally acquire genes or genetic elements by horizontal transfer, we know that they acquire new genes by the mechanism of gene duplication and divergence:  new genes arise by duplication of old ones, and then the functions of these once-identical genes diverge as they acquire new mutations.   Or, new genes can arise by unequal crossing-over between different genes, so that new genes arise by mixing bits of old ones.  Behe would count both of these as type 2 mutations (“gain of FCT”).  Think of all the genes that have arisen in eukaryotes in this way and gained novel function:  classic examples include genes of the immune system, Hox gene families, olfactory genes, and the globin genes.  And in many cases the origin of new genes via duplication or swapping of bits is untraceable because the genes originated so long ago and have diverged so greatly in sequence that their origin is obscure.

Vertebrates are thought to be the product of two whole-genome duplication events, giving rise to many genes with novel functions.  This has probably happened in yeast at least once, and many plants are the results of ancient “polyploidy” events in which entire genomes were duplicated at least once.  More than 40% of the genes in the human genome arose via gene duplications; this rises to more than 75% if we count those ancient rounds of whole-genome duplication. And over a third of the genes in the invertebrate  Drosophila genome arose via duplication, with most of these having new functions. There are many, many papers describing and discussing the importance of duplicated genes (and regulatory elements) as a source of evolutionary novelty; see, for example, Long et al. (2003), Wray et al. (2003), and Kaessmann et al. (2009).

While Behe’s study is useful in summarizing how adaptive evolution has operated over the short term in bacteria and viruses in the lab, it’s far less useful in summarizing how evolution has happened over the longer term in bacteria or viruses in nature—or in eukaryotes in nature.  In this sense it says nothing about whether new genes and gene functions have been important in the evolution of life.  Granted, Behe doesn’t make such a sweeping statement—his paper wouldn’t have been published if he had—but there’s no doubt that his intelligent-design acolytes will use the paper in this way.

Finally, this paper gives ID advocates no reason to crow that a peer-reviewed paper supporting intelligent design has finally appeared in the scientific literature.  The paper says absolutely nothing—zilch—that supports any contention of ID “theory.”


Behe, M. 2010. Experimental evolution, loss-of-function mutations, and “the first rule of adaptive evolution. Quart. Rev. Biol. 85:419-445.Kaessmann, H., N. Vinckenbosch, and M. Y. Long. 2009. RNA-based gene duplication: mechanistic and evolutionary insights. Nature Reviews Genetics 10:19-31.

Kaessmann, H., N. Vinckenbosch, and M. Y. Long. 2009. RNA-based gene duplication: mechanistic and evolutionary insights. Nature Reviews Genetics 10:19-31.

Long, M., E. Betran, K. Thornton, and W. Wang. 2003. The origin of new genes: Glimpses from the young and old. Nature Reviews Genetics 4:865-875.

Wray, G. A., M. W. Hahn, E. Abouheif, J. P. Balhoff, M. Pizer, M. V. Rockman, and L. A. Romano. 2003. The evolution of transcriptional regulation in eukaryotes. Molecular Biology and Evolution 20:1377-1419.

50 thoughts on “Behe’s new paper

    1. There does not seem to be a lot to rebut. I have yet to read it, but from this summary it appears to be a very cautious review that acknowledges its limits.

      Naturally, there will be those in the ID camp who will oversell this. They will be the ones who will need reminding about what Behe said and didn’t say.

  1. Even if Behe found an experiment that showed evidence of a Type 2 modification, no doubt he would have dismissed it as a carefully controlled labratory creation. Just as he critisized Miller-Uray type experiments in Darwins Black Box.

    Before I ever read that book, I had nagging suspicions about religion. Darwins Black Box opened my eyes to the depths people will go. In a way, I thank Mr Behe for that.

    1. As I understand it, nylonase resulted from two point mutations in an existing esterase enzyme without significantly reducing the original esterase activity.
      If so, that’s new function pretty cut & dried.

      1. Behe, bizarrely, limits his investigation to experimental cases, except, also bizarrely, for some random exceptions like sickle-cell hemoglobin.

        He totally ignores the ubiquitous evidence that gene duplication + change of function in one or both copies — he would probably say “that’s just sequence similarity” or some similar silliness.

        1. “He totally ignores the ubiquitous evidence that gene duplication + change of function in one or both copies — he would probably say “that’s just sequence similarity” or some similar silliness.”

          No he doesn’t. As is clear in the section Jerry quoted above, he considers these to be “Gain-of-FCT” mutations.

          1. He admits that, if they happened, they would be gain of function mutations, but he doesn’t acknowledge the massive evidence that they have, in fact, happened, and are the well-known and obvious source of most of the sort of novelty he is interested in.

            Instead, he puts a bunch of emphasis onto the fact that in a few mostly old, mostly short-term studies, he didn’t see many cases of such gain-of-function mutations, therefore they are rare, therefore (implied by Behe, explicit by the DI) evolution basically only works to make things simpler, not more complex. IMHO a competent review would have had to admit “of course, the fact that experimental studies have documented numerous examples of gene amplification under selection, and change of gene function under selection, this strongly supports the dominant scientific model for the origin of new genetic functions, which is gene duplication + modification”. But somehow he didn’t get around to saying such things…

      2. “Behe, bizarrely, limits his investigation to experimental cases, except, also bizarrely, for some random exceptions like sickle-cell hemoglobin.”

        Haven’t read the paper, but would doubt he intends it to make an ‘all inclusive’ case for anything; just points to consider.

        We see many argumental cases in fact, where a single event in nature is taken as justification for an evolutionary scenario. I don’t see in his abstract where he is doing likewise, emphasis mine:

        I show that by far the most common adaptive changes seen in those examples are due to the loss or modification of a pre-existing molecular function, and I discuss the possible reasons for the prominence of such mutations.

  2. Thanks for an objective review of the main points that Behe make, although I haven’t read the piece. Data is important, but not as important as objectivism in evaluating the data. We don’t always see this is reviews

    While Behe’s study is useful in summarizing how adaptive evolution has operated over the short term in bacteria and viruses in the lab, it’s far less useful in summarizing how evolution has happened over the longer term in bacteria or viruses in nature—or in eukaryotes in nature.

    Granted, but its emphasis I feel, is on disallowing a carte blanche reliance upon prokaryotic adaptations as confirmations of evolutionary processes in toto. I view evolutionary processes as categorically one, but in reality a myriad of mechanistic processes for multiple purposes, such as (broadly) adaptation, diversity, and to produce novelty. I put limits, however, on the third sub-category.

    If anything, the paper may strengthen the case for weakening the case for prokaryotic experiments (Lenski), and prokaryotic observations (nylonase) as verifications of evolution of a differing genera (phylogenetic novelty alterations). Or, it may not. But it may well stimulate further research efforts.

    1. Well, self assigned “Intelligent Design advocate”, that Behe leaves mechanisms out doesn’t mean that the process is “weaker” or can’t be observed but that it is richer and there is more to observe.

      We discuss science here, not creationist and inductionist religion, so define “weaker” as a meaningful characterization of hypotheses testing. Is 99 % certainty in rejection of a mechanism (say, creationism) “weaker” or “stronger” than 99 % certainty in rejection of said mechanism?

      If you can’t discuss the paper on the grounds of science, maybe you should find someone else to “bug” than a science blog. [Altogether now: “but this is not a blog!”]

      1. Also, I’m fairly certain that “Intelligent Design advocate” was ““Intelligent Design employee” a few years back – certainly under the Dover case!?

        If that is the case, it would be pertinent information too.

        1. Fair enough, in the presence of opinion and the absence of science. But I note that it has been known for a long time now that man’s ancestors were bipedal for a _very_ long time and many species back – so we don’t go up a tree.

          The creationist rock was overturned recently though.

  3. He didn’t? Oh, he did! I wondered why the DI could leave the Lenski results alone, when it annoyed them so. The answer is now here; of course they couldn’t.

    I’m no biologist, but I wonder how Lenski’s as I understand it “potentiating” mutations could be classified as either “loss-of-FCT” or “modification-of-function”. The citrate metabolism acquisition wouldn’t have happen unless an earlier mutation occurred in that particular lineage:

    Lenski’s experiment is also yet another poke in the eye for anti-evolutionists, notes Jerry Coyne, an evolutionary biologist at the University of Chicago. “The thing I like most is it says you can get these complex traits evolving by a combination of unlikely events,” he says. “That’s just what creationists say can’t happen.”

    Well, if that is the most damning he can come up with in his old age, i.e. lab # nature, I say good riddance. (Especially since that is what biologists have told us up front the whole time anyway.)

  4. I’d just like to point out that nuclear fusion was not observed in those experiments either but so what? It’s not as if any sensible person was expecting it. (Aha! No one expects the Spanish Inquisition!) Behe may as well write that a dog was not observe to transform into a monkey overnight or even in 10 generations of dogs. Standards for publication must be really low these days, especially if an author can claim a “law” via absence of evidence in a very limited case.

    1. Sorry – that was a “rule” rather than a law – but the rule is utterly meaningless anyway and should have been edited out.

  5. You get right down to it, it’s all micro evolution. But when you’re talking about thousands, possibly millions (even billions in the right circumstance), of generations, microchanges can build up to macroresults.

    1. Not to mention the absurdity of evolution proceeding primarily from the loss of FCTs – we’d need a god who created incredibly complex animals at the start. But once again, data suggest that it is not the case.

  6. How very odd. In the Q&A of Behe’s recent lecture in Glasgow, which I attended, a guy who was most likely a creationist asked him ‘do you know of ANY genetic mutations which have produced new information?’. Behe replied with ‘nope, there aren’t any as far as I know, apart from MAYBE genome duplication’. How does that answer square with the statement in his paper that:

    ‘It is certainly true that, over the long course of history, many critical gain-of-FCT events occurred.’

    It appears Prof. Behe is talking out of both sides of his mouth. To be expected, perhaps?

    1. Well, maybe not. Depends how the question was worded. I think he believes that gain-of-FCTs are due to God, not random mutations, and so if the question referred to *random* mutations, then his statements could be compatible.

      I think Behe pretty much accepts the time line and results of evolution (e.g., common descent, the accuracy of the fossil record, the conclusion from molecular data), but he just believes that God is/was necessary all along to step in and make those changes at the time they occurred. (And it kind of a safe position, because even if a laboratory experiment shows a gain-of-FCT, no reason he can’t invoke God having done it. After all, God can do anything.)

      1. And god didn’t do it in the experiments surveyed because he doesn’t like being watched by the filthy humans he loves so much.

      2. “Well, maybe not. Depends how the question was worded. I think he believes that gain-of-FCTs are due to God, not random mutations, and so if the question referred to *random* mutations, then his statements could be compatible.”

        This is exactly right — Behe is often slippery like this. Critics have often been led astray by Behe’s tactical ambiguities, as Boudry et al. documents in the same QRB issue…

  7. There may be a simple reason why gain of function mutations are difficult to ‘catch’ in prokaryotic laboratory experiments. If we take the analogous situation reagarding eukaryotes we can ask the question “how often can you catch a gene duplication event occurring?” The answer to that – as shown by modern genome analysis techniques – is apparently very often indeed. It is far more common that nearly everyone expected a few years back. Any single human individual will have some gene duplication event compared to another random human.
    If gene duplication events, or something similar, are the basis for most gain of function ‘mutations’ then we should ask when does the analogous scenario arise in prokaryotes. The answer to that is when they are mixed with bacteria of different strains or even different species that allow ‘sharing’ of genomic DNA or plasmids. This scenario is, naturally, something that tends to be avoided in most laboratories since it adds variables to the experiments that are very difficult to control. For instance Lenski’s experiments would be almost impossible to interpret if he allowed numerous other bacterial strains and species to interact with his original stock or their descendents.

  8. Pardon my lack of knowledge on the subject, but from this explanation it seems that bacteria and viruses don’t undergo gene duplication? Why is that?

    1. See Carl Zimmer’s comment in my latest post on this topic above. Apparently they do—at least in E. coli. It was my impression, though I may be wrong, that gene duplication is more common in eukaryotes than prokaryotes. I’m trying to find out more about this.

      1. It is certainly much more common in eukaryotes – at least if we are comparing the rates of gene duplication in the common genomic workhorses(!) E.coli and H.sapiens.

  9. So what do you guys have in the way of constructive mutations?

    Is there any evidence that genetic mistakes can accumulate in such a way as to give rise to novel protein machinery, novel body parts and novel body plans?

    Do you guys have any positive evidence for your position?

    For example:

    1- How can we test the premise that the bacterial flagellum evolved in a population that never had one via an accumulation of genetic accidents?

    2- How can we test the premise that fish evolved into land animals via an accumulation of genetic accidents?

    3- How can we test the premise that reptiles evolved into mammals via an accumulation of genetic accidents?

    1. I would suggest starting with Coyne’s book and going from there. There are multiple lines of evidence demonstrating shared common descent in cases 2 and 3, including step-wise changes in expressed traits such as limbs and eyes. The most likely mechanism for this is generation of small *mutations* (the scientific term – rather than the loaded term “genetic mistakes” that you use) and the weeding-out effect of selection. Now exactly how this works on a micro level is a detail that’s still in the process of discovery, and one that Behe’s paper, if taken in good faith, is a contribution to.

      Also importantly, what’s your alternative hypothesis, and how well does that hold up against falsification?

  10. The fundamental problem with this sort of analysis is that he’s arguing that beneficial mutations are rare, but neglecting that (regardless of rarity) any trait with survival benefit tends to propagate exponentially with time (and with degree of benefit, for non-zero benefits).

  11. Pingback: Quora

Leave a Reply