Embarassingly short of IDEAs

The IDEA Center is an organization that aims to support the diffusion of Intelligent Design (ID) notions by coordinating the activity of local chapters of (mostly) high school and undergraduate students (“IDEA clubs”) devoted to the study, discussion and promotion of ID.  Their spiffy new web site provides a wealth of material and information for club members and other ID-interested parties.  Their approach is very serious, scientific and professional.  They prominently announce:

Our mission statement states that the IDEA Center aims to:
1. Promote, as a scientific theory, the idea that life was designed by an Intelligent Designer
2. Educate people about scientific problems with purely natural explanations for the origins and evolution of life
3. Challenge the philosophical assumptions of Darwinism, naturalism, and materialism
From: The Science of IDEA

and assure:

This scientific approach [based on empirical observations of the natural world - AB] is the method that the IDEA Center takes when discussing intelligent design theory.
From: Religious and scientific affiliations

In the spirit of scientific inquiry, the site provides a number of FAQs and Primers related to evolutionary biology subjects, ID arguments, etc.  Today I am just going to focus on their recent and particularly bad “Featured FAQ” “Can irreducibly complexity be evolved via gene duplication and co-optation of parts?” (hereafter referred to as “IC, duplication and co-optation”, for short) which was brought to the attention of Panda’s Thumb members through the ID discussion board at ARN.

This FAQ is meant to address whether non-direct evolutionary pathways can generate structures with the characteristics of “irreducible complexity”, via gene duplication followed by the recruitment (co-optation) into the system of proteins or protein complexes of different function encoded by the duplicated genes.  The author seems to have no doubts about what the answer to the question is:

The short answer: No. Darwinian evolution requires that a biological structure be functional along every small step of its evolution. Irreducibly complex structures fail to perform any useful function if changed in any way, and thus cannot be built via the Darwinian mechanism. Thus, irreducibly complex structures are in principle unevolvable via Darwinian evolution. [emphasis mine - AB] If a biological structure can be evolved (via co-optation of parts or otherwise), it is not irreducibly complex.
From IC, duplication and co-optation

Let’s ignore the obvious erroneous claim that Darwinian mechanisms require that every step in an evolutionary pathway be selectable, and the tautological reasoning (“An irreducibly complex structure cannot evolve; if it can evolve, it’s not irreducibly complex.”).  The author(s) here adopt Behe’s original “slam-dunk” definition of IC, which implies that IC fulfills Darwin’s famous challenge that a complex organ which “could not possibly have been formed by numerous, successive, slight modifications” would invalidate his theory.  However, it is well known that Behe later modified this claim, turning IC into a mere probabilistic, quantitative argument.  The author(s) are aware of this, since the FAQ later on quotes Behe’s own original statement reformulating the IC concept in probabilistic terms:

“An irreducibly complex evolutionary pathway is one that contains one or more unselected steps (that is, one or more necessary-but-unselected mutations). The degree of irreducible complexity is the number of unselected steps in the pathway.”
(A Response to Critics of Darwin’s Black Box, by Michael  Behe, PCID, Volume 1.1, January February March, 2002; iscid.org/)
From: IC, duplication and co-optation

Not only it is conceded here that irreducible complexity can, in principle, evolve through non-selected intermediates, contradicting the “short answer” above, but in fact, if one adheres to this new definition, it would be surprising if many evolved complex systems were not IC to at least some degree, neutral mutations and drift being so pervasive in evolution.  So, this FAQ starts already on a bad footing, with lapses of logic as well as crucial and self-contradictory misunderstandings of one of ID’s own founding concepts.  And it only gets worse from there.

The next step in the FAQ argument is a criticism of Russell Doolittle’s explanation for the evolutionary origin of one of Behe’s original examples of IC system, the clotting cascade.  The FAQ author’s judgment of Doolittle’s argument, supposedly presented at a graduate student seminar, is scathing:

The essence of Doolittle’s response was to proclaim nothing more than homology between various proteins in the blood clotting cascade…  This was the essence of Doolittle’s response to Behe: the blood clotting cascade proteins are similar (to varying degrees), therefore they must have evolved from one-another via duplication and co-optation. This mere sequence similarity, nothing less, and nothing more, formed the basis for Doolittle’s belief that the irreducibly complex blood clotting cascade could evolve. There was no discussion of intermediate stages showing how stepwise modifications in the current pathway could have been reversed back to a more simple stage. This is what graduate students were taught in a seminar directly intended to rebut the claims of Behe. From: IC, duplication and co-optation

As an explanation of Doolittle’s reasoning, the FAQ shows a figure that looks essentially like this one on the left, minus my disclaimer (you can find the original in the FAQ, until they change it), accompanied by the following description:

Above we see 5 proteins involved in the blood clotting cascade, arranged in a tree simply according to their percent sequence similarity.
From: IC, duplication and co-optation

I wasn’t at Doolittle’s seminar, but I have every reason to believe that the FAQ author wasn’t either, or if he/she was, he/she wasn’t paying much attention, because the figure shows a phylogenetic tree of hemoglobins, which are proteins involved in oxygen and CO2 transport, and have absolutely nothing to do with blood clotting.  There are also 5 other errors (!) in the figure (2 obvious, 3 less so), but I’ll leave it to the readers to identify them (first one to get them all right wins a virtual pint at the PT virtual pub).

Now, Doolittle’s argument may have sounded unconvincing to an ID faithful, but two things are almost certain: he didn’t claim that hemoglobins are involved in blood clotting, and his argument did not rest simply on protein similarity.  Several reviews of the evolution of the blood clotting cascade, including one by Doolittle himself, have appeared in the last couple of years - regardless of the original Doolittle seminar contents, certainly enough time for the FAQ author to get acquainted with the data [1].  In fact, the blood clotting cascade has been essentially ignored in recent ID literature, perhaps because of the accumulating evidence for evolutionary precursors and overall evolutionary plasticity of the clotting system in various vertebrates, as discussed in the references above.  Recent evidence has linked the appearance of some duplicated gene pairs (those for clotting factors IX and X, V and VIII, respectively) to a global genome duplication that is thought to have occurred right before the emergence of jawed vertebrates (sharks, bony fish, and land vertebrates).  If this is correct, jawless vertebrates (or agnathans, like lampreys and hagfish) would be expected to display simplified cascades containing a reduced set of these components.  Indeed, although jawless fish gene data are still partial, of the major clotting components only thrombin, fibrinogen, and 3 members of the FVII/FIX/FX/protein C family have been identified so far.  The origin of the FV/FVIII duplicated pair is also thought to likely date to the same whole genome duplication, and the search for their agnathan precursor is ongoing.  As the jawless fish gene catalog becomes more complete, it will be interesting to see whether these predictions will be borne out. 

Let’s move on.  The next target of the FAQ is Nick Matzke’s recent paper on the evolution of the bacterial flagellum.  This is a dense 30-something page-long, thoroughly researched and referenced work, and although it may be hard going for the lay reader, I strongly recommend giving it a try to understand the level of analysis and inferences associated with an evolutionary approach (and compare that, if you will, to any IDEA FAQ, or any of the essays on Dembski’s web site).  Nick’s paper identifies individual subcomponents in the bacterial flagellum, provides evidence for their relationship to non-flagellar proteins based on structural and functional considerations, and proposes a testable model that defines several potential major intermediates on a hypothetical evolutionary path from a non-flagellar system to motile flagella. Note that a central prediction of evolutionary theory is that a new system/protein will display structural and functional similarities with other systems/proteins related to it by descent.  Furthermore, the degree of such similarities (homologies) will be in general inversely proportional to the time elapsed since the evolutionary separation: the more ancient the separation, the weaker/more scarce the homologies.  On the other hand, ID makes no prediction at all in this respect: the Designer could design from scratch, or tweak existing components, or transfer them unchanged, entirely whimsically. 

So, one may expect ID advocates to carefully read Nick’s paper, admit there is more evidence for the evolution of the flagellum they had initially anticipated, argue if possible with individual pieces of such evidence, and perhaps even start a research program to test Nick’s hypothesis in their own lab, in the hope of proving it wrong…. yeah, right.  The FAQ’s author response, following in the footsteps of an earlier mindless hatchet job by Dembski, consists in scornfully denying that any such evidence exist, at least not to the extent that would be necessary to convince them:

Without getting into further detail, the main substance of the rest of Matzke’s explanation for the origin of the flagellum is to look at current flagellum parts, and find homology to other parts, and simply claim that they were duplicated and co-opted. The detail barely goes beyond these sorts of claims.
From: IC, duplication and co-optation

So, Nick presents precisely the kind of evidence that evolutionary theory would predict  to find for the evolution of an ancient system like the flagellum, but the response is that the evidence that ID advocates really want to see is of the kind that evolutionary theory would predict is going to be almost impossible to find.  Oh, well…

Had enough already? The best has yet to come!  Most if not all hypotheses on the possible evolution of IC systems, like Nick’s, rely on gene duplication and co-option events.  Since the the evidence for gene duplication in evolution is so pervasive that even for the most devoted ID advocate cannot just pretend it doesn’t exist, this FAQ’s approach is to instead deny that duplications can possibly do the job:

Duplicating a gene doesn’t help the evolution of complexity much, because once you duplicate a gene, you have a new piece of genetic information to play around with, but what good is that to you?
From IC, duplication and co-optation

Now, the thoughtful reader may think that “a new piece of genetic information to play around with” is exactly the substrate evolution requires to evolve new functions, but apparently this isn’t the case for the FAQ author. 

To support the claim of “helpless” gene duplications, the FAQ then plunges into a discussion of an actual scientific article published 4 years ago in Science by Michael Lynch and John Conery, and that’s when the bottom really falls off.  The FAQ says:

Lynch and Conery found that the average gene duplicates about once every 100 million years  (Lynch, M., Conery, J. S., “The Evolutionary Fate and Consequence of Duplicate Genes” Science 290:1151-1155  (Nov 10, 2000)) —that’s pretty rare. If the gene you need duplicates once every 100,000,000 generations, that doesn’t give you a very good chance of getting it when you need it. Furthermore, it has been found that, “the vast majority of gene duplicates are silenced within a few million years, with the few survivors subsequently experiencing strong purifying selection.” (Lynch and Conery, 2000).
From IC, duplication and co-optation

Never mind the confusion between generations and years, the main issue here is that Lynch and Conery’s point is exactly the opposite, and explicitly so.  The very abstract of the paper, from which the FAQ author took his quote, also says:

Gene duplication has generally been viewed as a necessary source of material for the origin of evolutionary novelties, but it is unclear how often gene duplicates arise and how frequently they evolve new functions. Observations from the genomic databases for several eukaryotic species suggest that duplicate genes arise at a very high rate, on average 0.01 per gene per million years.[emphasis mine -AB]
(Lynch and Conery, 2000)

Lynch and Conery in fact argue that genes are continuously duplicated and exposed to selective forces at an unexpectedly high rate (as high as that of conventional point mutations on a per nucleotide basis!), and that although most newly duplicated genes are rapidly lost,

…even in the absence of direct amplification of entire genomes (polyploidization), gene duplication has the potential to generate substantial molecular substrate for the generation of evolutionary novelties.
(Lynch and Conery, 2000)

The FAQ author also seems to believe that the figure of one duplication every hundred million years for any one gene represents the rate of initial duplication events, and deems this frequency “pretty rare”.  However, that estimate was in fact derived by calculating the number of very recently diverged duplicates (those with a minimal number of silent nucleotide substitutions) in the human, Drosophila and C. elegans genome databases.  Since the vast majority of gene duplications are rapidly lost before they can spread to many individuals, this subset consists therefore only of the duplicates which, through positive selection or (rarely) neutral drift, have become common enough in the population to be present in genome databases.  To put things into perspective, this means that something like 6% of our genes (that’s about 2,000 genes, at last count) must have generated common duplicates at some time point since humans have diverged from chimpanzees. Looking forward instead of backwards, Lynch and Conery provide this estimate:

Given this range, 50% of all the genes in a genome are expected to duplicate and increase to high frequency at least once on time scales between 35 to 350 million years.
(Lynch and Conery, 2000)

That seems like a heck of a lot of gene duplications, and abundant fodder for evolution, if you ask me. [2] 

How the FAQ author can have quoted from this paper the few passages that suited his/her interpretation, ignored several prominent unfavorable ones, and completely misunderstood the paper’s actual meaning, frankly boggles the mind. 

Indeed, the author has chosen the wrong time to argue against the generation of evolutionary novelty through gene duplication, as the unprecedented amount of evidence derived from the fast accumulation of genomic sequence information from different organisms has become so significant and generally useful that the National Center for Biotechnology Information, which manages the information databases for the NIH, found it appropriate to devote specific resources to the phylogenetic analysis of gene homology clusters.  Direct whole-genome comparisons are shedding light on large-scale duplication events, like this one in yeast, recently described in Nature, which help track the fates of new genes, and learn about their origin and diversification processes.  As the FAQ author grumbles against “unlikely miracle mechanisms”, biologists continue to mine the actual evidence. 

I will skip commenting on the last few paragraphs, which include a bizarrely contrived example involving - I kid you not - a lock, a key, a hammer and what looks to me like a gin trap (though the FAQ author identifies it as a vice) which purportedly illustrates how evolutionary co-optation is utterly impossible, and of course the ritual homage to Dembski’s Explanatory Filter.

The extensive errors in basic biology, half-baked arguments, as well as the misunderstandings and/or misrepresentations of scientific evidence and claims in this FAQ make a mockery of the pompous pronouncements of the IDEA Center about their goals and adherence to scientific standards.  Quite a lesson for the “future generations of ID scientists”! 

Notes and references:
[1] Krem MM, Cera ED. Evolution of enzyme cascades from embryonic development to blood coagulation. Trends Biochem Sci. 2002, 27:67-74.
Davidson CJ, Hirt RP, Lal K, Snell P, Elgar G, Tuddenham EG, McVey JH. Molecular evolution of the vertebrate blood coagulation network. Thromb Haemost. 2003, 89:420-8.
Jiang Y, Doolittle RF. The evolution of vertebrate blood coagulation as viewed from a comparison of puffer fish and sea squirt genomes. Proc Natl Acad Sci U S A. 2003, 100:7527-32.

[2]Incidentally, my own experience in the study of gene duplications dates back to my undergraduate and graduate years, when I worked on duplication/deletion mechanisms in a family of human genes involved in antibody formation.  Duplications and their rarer mechanistic counterparts, deletions, were found to be so common in humans that over time we realized that what was originally considered the “normal” locus structure actually represents less than 50% of the existing alleles in some populations.  In fact, duplications are so prevalent that they have had time to become duplicated themselves, giving rise to triplicated loci!  Granted, this is an extreme example (multigene families are notoriously unstable, and were therefore excluded from Lynch and Conery’s analysis - which in this respect underestimates the overall rate of gene duplication), but I think it nicely conveys the scale of new gene formation and genome dynamycs in general.