t f p g+ YouTube icon

Is There an Edge to Evolution? Part 2

Bookmark and Share

October 18, 2010 Tags: Design
Is There an Edge to Evolution? Part 2

Today's entry was written by David Ussery. Please note the views expressed here are those of the author, not necessarily of BioLogos. You can read more about what we believe here.

An Analysis of Michael Behe’s book, The Edge of Evolution

In his previous post, Ussery discussed his personal reasons for being interested in “The Edge of Evolution.” He went on to discuss two aspects of the book he appreciates, and he showed that he and Behe are in agreement that all living organisms have arisen through common descent from a single ancestral species.

In this post, however, Ussery says that Behe has presented a vastly over-simplified view of what scientists know about the origin of genetic diversity in the history of life. Here is his analysis based on Chapters Two and Three.

The accompanying figure illustrates the amount of genetic diversity in the bacterial world and beyond. It shows that even a single species of bacteria (E. coli) contains a vast reservoir of different genes. The term “orthologous” refers to genes in different species that clearly resemble one another and are thereby believed to have a common ancestry. Genes which are not orthologous are found in the species on the right side but not on the comparator species on the left.

Chapter 2 - Arms Race or Trench Warfare?

This chapter is about one of the classic examples of evolution: malaria and sickle cell anemia in humans. Behe observes (correctly, in my opinion) that the mutations that are responsible for helping some humans fight malaria are bad mutations. 'The first point is that the two examples he cites, sickle and Hemoglobin C (HbC), (two mutations that help the body resist malaria), are quintessentially hurtful mutations because they diminish the functioning of the human body. A second point is that “the mutations are not in the process of joining to build a more complex, interactive biochemical system.” (page 34).

Fair enough—and it is well known that harmful mutations, in the sense of wrecking something or making a pathway not work, occur much more frequently than beneficial mutations. However, Behe goes on to claim that there are “absolutely no studies' to document a molecular basis for the “coherent development of a single trait in a Darwinian arms race.” But this is highly erroneous . True, the example he gives us is not a “good mutation” - but to just blatantly claim that nothing has been done is showing his ignorance of the literature.

For example, consider this from the abstract of a recent review article, with the title “Origins, evolution, and phenotypic impact of new genes,” published in Genome Research. “The array of mechanisms underlying the origin of new genes is compelling, extending way beyond the traditionally well-studied source of gene duplication. Thus, it was shown that novel genes also regularly arose from messenger RNAs of ancestral genes, protein-coding genes metamorphosed into new RNA genes, genomic parasites were co-opted as new genes, and that both protein and RNA genes were composed from scratch (i.e., from previously non- functional sequences).” This is a new article, but many of the references in this article date to long before The Edge of Evolution was written, and some even date to before Darwin's Black Box was published, more than a decade ago.

Then there's another article about recent evolution of beneficial mutations in humans. There are many, many articles published on this sort of idea, and to claim that not a single study has been done is essentially a play on the ignorance of the readers! It is as if the hope is that the readers are ignorant of the scientific literature, and either too lazy or not competent to have a look through PubMed and see what is really out there.

Chapter 3 - The Mathematical Limits of Darwinism

One of my Ph.D. students was a mathematician, and I can still remember trying to read through his paper—lots of formulas—and sometimes they were difficult for me to understand. I have since learned that many people in math departments have a strong disliking for statisticians - I used to naively think that the two are the same. In this chapter, it looks as though Behe has confused mathematics (in the title) with statistics (what is actually discussed in the chapter). What's worse, the numbers he uses are based on bad assumptions, and are way off from what is known in the field by experimentalists doing current research in this area. Thus, unfortunately, his conclusions are not as strong as they might seem at first glance.

First, in calculating the odds of a single mutation in a protein, one has to take into account the chances of a mutation in the DNA sequence, because this is where mutations happen in biology - this is part of the 'central dogma' of molecular biology - that the information flows from DNA to RNA to protein, but not from proteins back to DNA. Thus, if a protein has a particular amino acid changed, this can be traced back to a change in the DNA sequence. Behe says ”resistance to chloroquine has appeared fewer than ten times in the whole world in the past century” - but what is meant by this shorthand is that we have documented evidence of this happening only a few times - that's not the same as knowing definitively that this HAS happened only those few times. Lots of things [like mutations leading to drug resistance] happen all the time that don't get seen and documented.

Then, based on this vastly over-simplified estimate, he suggests that the odds of a parasite developing resistance to chloroquine is one in 1020, whilst the odds of developing resistance to another drug (atovaquone) is one in 1012. Since the former, he says, involves two amino acid changes, while the latter involves on one, from these two numbers, it is concluded that the chance of having mutations which change two amino acids in a protein is a hundred million times lower (10-20 vs 10-12) than that for just getting one.

But this just simply does not make sense. Even within E. coli, the well known work-horse of molecular biology, take the order of amino acids in any one of its 5000 or so proteins, and compare that arrangement to that in other E. coli strains and you will find LOTS of differences. For many proteins in E. coli, the level of identity between strains is around 80% - that is, about twenty out of every hundred amino acids are different - so to say that the odds for a double mutation (2 amino acid changes out of 100), is essentially impossible, when we observe 10 times that amount of diversity (20 differences for every 100 amino acids) in natural populations is speaking from ignorance. We see ten times the number of changes which Behe says is almost impossible all around us within a single species without even the need to generate new mutations.

I’ll discuss the vast differences found with various sequenced E. coli genomes later, but getting back to this chapter and the mutations in the two different spots within a single gene, Behe concludes:

On average, for humans to achieve a mutation like this by chance, we would need to wait a hundred million times ten million years. Since this is many times the age of the universe, it's reasonable to conclude the following: No mutation that is the same complexity of chloroquine resistance in malaria arose by Darwinian evolution in the line leading to humans in the past ten million years. (page 61, emphasis in the original).

But again, if one takes a deep breath, and looks at what is known, the mutational frequency that we can actually measure in humans is many times greater than that upon which Behe's assumptions are based. His argument is that the chances of getting useful mutations at two sites in the same gene are highly unlikely. But can we assess how likely mutations, which are likely to change the function of a gene, occur? One of the underlying assumptions of molecular biology is that sequence determines structure, and that this structure determines function. Hence, a major structural change is likely to have a different function. So how common are mutations that result in structural changes in proteins?

Surprisingly Common! One out of every 21 births in humans have some sort of STRUCTURAL change (and hence likely a functional change) in a protein, just from insertions from a single transposable element (alu), common in humans. It is already evident that Behe has a real problem with “random” mutations – but I think perhaps he is confusing ‘randomness’ with ‘purposelessness’.1 More about that in my next post.


1. I think many people don’t really understand randomness - for more on this see David Bartholomew’s excellent book “God, Chance and Purpose - Can God Have it Both Ways?” (Cambridge University Press, 2008), and also my “Purpose-Drive iPod” essay (Christian Century, 23 September, 2008, pages 11-12).

David Ussery is an associate professor of comparative microbial genomics at the Center for Biological Sequence Analysis at the Technical University of Denmark and on the faculty at the University in Oslo, Norway. Ussery is the co-author of Computing for Comparative Microbial Genomics and has authored or co-authored 130 articles for science and professional journals. He is also a frequent public speaker on the topic of bacterial genomics.

< Previous post in series Next post in series >

View the archived discussion of this post

This article is now closed for new comments. The archived comments are shown below.

Page 5 of 9   « 2 3 4 5 6 7 8 »
Dave Ussery - #36004

October 23rd 2010

Behe’s h-index is 16 - he has one paper from 1991, that has been cited more than a hundred times.  Not bad, but if we consider when he got his Ph.D. (1978), then we get a ratio of 0.5 - which means he would not qualify for consideration as a faculty member in my group, because he has simply published too little….

sy - #36074

October 24th 2010


The h index has been criticized as a method for productivity comparison, because it never goes down with time. Your approach solves that problem, but creates another, in that the denominator will generally rise at a faster rate than the numerator. Your example of Behe is a good one, In order for him to be hired on your faculty, he would need an h index of 32, pretty rare.

My own h index (if anyone cares) is 20, but I am in the same age boat as Behe, and would also not qualify.

Of course if we are talking about fairly junior people, than my caveat vanishes, since at early stage, and low h index, one should get a slope of about 1 when plotting h index against time. But it isnt linear at longer time scales.

Rich - #36101

October 24th 2010

Arthur Hunt:

“What I have in mind, to give an example, is the repeated reference to the value 10^-20 as the frequency of occurrence of a “CCC”.  However, even a cursory search of the literature reveals that functionally analogous mutational events may occur with rather different frequencies; off the top of my head, the range I can cite is between 10^-8 and 10^-20.”

Fine.  Let’s take the lower figure you’ve found in the literature.  That’s a probability of 1 in a hundred million.  Now, to get from an artiodactyl to a whale, you’ve got to have many CCCs.  And they all have to happen in about 9 million years.  So the probability of the transition’s occurring, if it occurs by purely neo-Darwinian means, then becomes the product of a string of 1 in a hundred million probabilities multiplied by each other—a smaller number in relative terms, but still unimaginably large in absolute terms.  Even if you reduce the exponent to 6 instead of 8, you still have a colossal number.

This is the problem with so many ID critics:  they are so focused on technical minutiae that they can’t see the bigger picture.  In the bigger picture, Behe’s number could be out by several orders of magnitude; neo-Darwinism still has big problems.

Dave Ussery - #36137

October 24th 2010

sy - #36074 - h-index

Dave Ussery - #36139

October 24th 2010

sy - #36074 - h-index

sorry - my computer slipped and I would up ‘submitting’ too soon!  Sy, I agree with you on this - there are OF COURSE other things to take into consideration.  But I think it’s a worthy goal, especially for young post-docs and assistant professors to try and publish at least one well-cited paper annually.

I forgot to mention that these numbers for the h-index are only available from a company, and most universities pay a subscription fee to access them, so unfortunately it’s not easily available for the public.  But I think “google scholar” has something different.

Arthur Hunt - #36202

October 24th 2010

Rich:  Now, to get from an artiodactyl to a whale, you’ve got to have many CCCs.:

Prove it.  For example, provide some quotes from some accomplished ID-friendly biochemist that shows, with references to the peer-reviewed research literature, that a “CCC” (ANY “CCC” - heck just one would do) must have been required, to the exclusion of any of the numerous other molecular mechanisms that underly developmental processes, to produce a different between an artiodactyl and a whale.  Prove that mechanisms* such as recruitment of other transcription factors (either extant or the products of duplication events) that need not involve simultaneous multiple mutations, remodeling of developmental pathways via gain or loss of microRNA targeting sequences in mRNAs that encode regulatory proteins, remodeling of developmental pathways via gain or loss of Ub E3 ligase-mediated targeting of strategic regulatory proteins (neither of which would involve a Behe-like “CCC”), or changes in the timing of synthesis or degradation of a regulatory signal (such as, say, a retinoic acid-related metabolite), cannot possibly have been involved in any known or hypothetical evolutionary transition

Arthur Hunt - #36203

October 24th 2010

*These latter mechanisms are in fact known to play roles in developmental and evolutionary transitions.  CCC"s, OTOH, not so much so.

Arthur Hunt - #36205

October 24th 2010

One request, Rich - don’t go moving goalposts.  You raised the issue of CCC’s in the context of artiodactyls -> whales, it’s up to you to support your assertion with facts.

And don’t resort to the excuse that you don’t buy the hypothetical series of evolutionary events.  You, or Behe, or Wells, or Nelson, or Meyer, or Axe, or Sternberg all have access to the literature and should easily be able to identify important developmental differences AND their underlying molecular foundations.  One needn’t accept that this evolutionary progression is true; one need only study the development of the individual organisms and identify the molecular changes that must have occurred if evolution were true.  That’s the sort of information that you should be able to draw upon to support your assertion.

Bilbo - #36210

October 24th 2010

Dave;  “Are you REALLY disappointed to learn that I say the Nicene Crede, that I believe in Jesus Christ as the son of God?

Yes I am disappointed.  First, it means that you will be held to a higher standard than an atheist, at the last judgment.  You will be asked why you, a supposed Christian, thought it was all right to lie about someone else.  You’re replying that you can say the Nicene Creed will avail you little comfort.  Second, it means that you are making the rest of us Christians look bad.

My “misquotes” of Behe are straight from the text - I try and include the page number.

REALLY???  You wrote:
Behe goes on to claim that there are “absolutely no studies’ to document a molecular basis for the “coherent development of a single trait in a Darwinian arms race.””  (no page number)

Where is that in the text?  The correct answer is that it isn’t in the text.


Bilbo - #36211

October 24th 2010

What is in the text (as I already pointed out at #35254):

There is no reason to expect the coherent development of a single trait in a Darwinian arms race….Although there have been some studies showing modest arms races with smaller animals—ants, other invertebrates, and microorganisms—there have been no studies that document that large animals change in the way Dawkins supposes.”  (p.42)

So you misquoted Behe.

You wrote:

  “Then, based on this vastly over-simplified estimate, he suggests that the odds of a parasite developing resistance to chloroquine is one in 1020, whilst the odds of developing resistance to another drug (atovaquone) is one in 1012.”  (again, no page number)

Where did Behe make this estimate?  The correct answer is that he didn’t make this estimate.


Bilbo - #36212

October 24th 2010

cont.  What Behe actually wrote (as I already pointed out at #35261):

”For instance, in the case of atovaquone, a clinical study showed that about one in a trillion [1012] cells had spontaneous resistance. [15]  Nicholas White of Mahidol University in Thailand points out that if you multiply the number of parasites in a person who is very ill with malaria times the number of people who get malaria per year time the number of years since the introduction of chloroquine, then you can estimate that the odds of a parasite developing resistance to chloroquine is roughly one in a hundred billion billion. [16]  In shorthand scientific notation, that’s one in 1020.” (p.57)

So you attributed to Behe something that other people did:  estimate the odds of developing resistance to atavaquone and chloroquine.

Bilbo - #36213

October 24th 2010

You wrote:  “Since the former, he says, involves two amino acid changes, while the latter involves on one, from these two numbers, it is concluded that the chance of having mutations which change two amino acids in a protein is a hundred million times lower (10-20 vs 10-12) than that for just getting one.”  (again, no page number)

Where did Behe write this?  The correct answer is Behe did not write this.  What Behe said was (as I already pointed out at #35266):

”The ratio of the two numbers shows that the malarial parasite is a hundred million times (108) less likely to develop resistance [not “mutations] ....”

Bilbo - #36214

October 24th 2010

David, if you were only misquoting Behe, or taking him out of context, or attributing to him things that others said, it might not be that great a sin.  But you did all of this in order to show how wrong Behe was about things.  And for that, you are storing up for yourself greater punishment.  I suggest you come clean, while you still can.

Bilbo - #36215

October 24th 2010

Finally, you ask me, “Um, so what exactly WOULD it take to convince you that perhaps everything Behe says is not true? Or is this possible?

Um, so what exactly WOULD it take to convince you, David, that perhaps Behe is right about the main things he says?  Or is this possible?

Rich - #36230

October 24th 2010

Arthur Hunt:

Don’t misdirect with paragraphs full of technical jargon.

You complained about the exponents in Behe’s equations being too high.  I said it didn’t matter if they were much lower, because when you multiply the probability of all the independent events that have to take place to turn an artiodactyl into a whale, you get a huge number even with lower exponents.  And if you want to drop CCCs, that’s fine.  You still have probabilities of, like, one in a million for any given point mutation (I’m sure you’ll correct me on the exact number), and you need thousands of specific beneficial mutations (when most are deleterious or useless) to do the job.  And you’re not dealing with bacteria, but mammals, with reproductive rates orders of magnitude lower than bacteria.  You’re up against a big problem.  But, like your colleagues, you refuse to provide hypothetical pathways and probability calculations to go with them.  Why?

Bilbo - #36231

October 24th 2010

BTW, David, I already linked to my evaluation of Behe’s edge, which will be far fairer than anything you have to say about it:

Bilbo - #35385

October 19th 2010

Hi Sy,

Let me cut this conversation short by referring you to a short evaluation of Behe’s edge by someone I admire very much:


Arthur Hunt - #36248

October 24th 2010

Arthur Hunt:

Don’t misdirect with paragraphs full of technical jargon.

LOL.  In other words, don’t try to confuse the issue with facts.

because when you multiply the probability of all the independent events that have to take place to turn an artiodactyl into a whale,

Just how many independent events are we talking about?  What does the ID vanguard (Behe, Meyer, Nelson, Wells et al.) say?  What research along these lines have they done?  What peer-reviewed literature to they cite when they make statements that pertain to this number?

Rich, I suspect that, as is the case with “CCC’s” and whale evolution, you are just making things up. 

You still have probabilities of, like, one in a million for any given point mutation (I’m sure you’ll correct me on the exact number)

Well, if you look into the abundant research done with the so-called Big Blue tool for measuring mutation rates in somatic and germ cells, you will find that this number is very, very low.  At the low end of the scale, it is safe to say that each and every new-born mammal will have at least 20* new mutations not found in either parent.

I’ll leave it to readers to figure out the ramifications.

Arthur Hunt - #36250

October 24th 2010

* I know, I know - informed readers will tell us that this number is far less than that suggested by most Big Blue studies.  But it is only a slight bit lower than other estimates I have seen, and I use it here to try and err as much as is reasonable on the side of the anti-evolutionists here.

Arthur Hunt - #36336

October 25th 2010

A point of clarification:  “this number” in 36248 referred to Rich’s “one in a million”.  Whether or not it is lower than the other studies I have in mind depends on just what “this number” is.  If this number is the per-nucleotide frequency of mutation during replication, then it actually translates to several thousand mutations per individual.  I suspect that this is not what Rich means.  Rich was rather cryptic about this, so I will try to speak about mutations per individual.

Sorry for any confusion.

Rich - #36455

October 25th 2010

Arthur Hunt:

Yes, I was referring to “the per-nucleotide frequency of mutation during replication.”

Several thousand mutations per individual.  Nolo contendere.  But do all of these bear on the form and function of the adult offspring?  Or are a lot of them in the portions of DNA that biologists have dismissed as “junk”?  You can’t have it both ways.  If, say, 90% of the DNA has no effect on the form or physiology or instincts of the adult creature, and if the mutations truly occur randomly along the genome, then 90% of those “thousands” won’t have any effect as far as evolution goes.  So you have to reduce your number proportionately.  Or do you dissent from your colleagues, and hold the ID view, i.e., that it is likely that much more of the DNA is functional than current biologists have been inclined to admit?

But let’s keep focused, Prof. Hunt.  Suppose we contemplate the transformation of a shrewlike creature into a bat, by means of random mutations.  Can you tell me which nucleotides have to be altered to give the shrewlike animal a bat’s wings?  A bat’s sonar?  We need this information for any calculation of the probability that the transformation could have occurred in X million years.
Where can I find it?

Page 5 of 9   « 2 3 4 5 6 7 8 »