<i>Discover Magazine</i> Fails With Miller's Failure To Refute Behe - Evolution News & Views

Evolution News and Views (ENV) provides original reporting and analysis about the debate over intelligent design and evolution, including breaking news about scientific research.

Evolution News and Views

Discover Magazine Fails With Miller's Failure To Refute Behe

This latest installment of my ongoing responses to Ken Miller regarding the irreducible complexity of the blood clotting cascade will critically analyze Professor Miller's citation of a 2008 paper co-authored by blood clotting expert Russell Doolittle. Citing to Doolittle, Miller claims that the lamprey lacks blood clotting components that Michael Behe, in Darwin's Black Box, actually did describe as being part of the irreducibly complex core of the blood clotting cascade. The problem for Miller is that Doolittle's conclusion was based on there allegedly being only one gene in the lamprey homologous to blood clotting factors V or VIII, but Doolittle's reported data belies that conclusion: it shows there were multiple potential homologues for those factors--including at least two conspicuous homologues that imply both factors V and VIII might be in the lamprey. Other than Doolittle's self-professed "hopes," his reported data offers no compelling reason to believe that the lamprey lacks either factor V or VIII of the blood clotting cascade. This means Miller's argument against Behe fails.

This current response to Miller recently got more interesting. The website of the anti-ID media outlet Discover Magazine is presently touting Dr. Miller's alleged refutation of me on the irreducible complexity of the blood clotting cascade as an "intelligent design fail." Apparently Discover has not been closely following this exchange of late. As I documented in a response to Miller last month (titled "Ken Miller's Only a Theory Misquotes Michael Behe on Irreducible Complexity of the Blood Clotting Cascade"), Miller was flat wrong to claim that the puffer fish's lack of certain blood clotting factors refuted any of Behe's arguments in Darwin's Black Box. A review of Darwin's Black Box--and Behe's clear explanation of own his writing--unambiguously demonstrates that in Darwin's Black Box, Behe did not argue that the factors in the intrinsic initiation pathway of the blood clotting cascade (which are the ones missing in the puffer fish) were irreducibly complex. And now today we see that the data reported in the Doolittle paper doesn't actually support the claims Miller wants to make.

Should Miller's fundamental and repeated misrepresentations of Behe in court, in lectures, and in his latest book qualify as a "fail"? How about Miller's reliance upon a paper that, when closely scrutinized, doesn't support the point he wants it to make? Is that a "fail"? How about Discover Magazine's devoted and forceful promotion of Miller's straw man attacks and misrepresentations of Behe's arguments? Read on and decide for yourself.

Miller Relies on Russell Doolittle
In his response to me on the irreducible complexity of the blood clotting cascade, Ken Miller claims that even if I am correct that he didn't accurately represent Behe's arguments about irreducible complexity during the Dover trial, Behe's arguments are still refuted by a much more recent paper. That paper was co-authored by Russell Doolittle in 2008. According to Professor Miller, Doolittle's paper "reports on a careful search through the lamprey genome. The lamprey, as luck would have it, has a perfectly functional clotting system, and it lacks not only the three factors missing in jawed fish, but also Factors IX and V."

Behe did not claim that Factor IX was part of the irreducibly complex core of the blood clotting cascade, but Behe did argue that Factor V was part of that core. Since the lamprey allegedly lacks Factor V, Miller claims that Behe's argument is refuted. But a closer analysis shows that Miller has over-stated the conclusions of Doolittle's paper, and that some of Doolittle's own conclusions are highly questionable given the data presented in the paper.

Doolittle's Basic Conclusions
As we saw in his statement quoted above, Miller effectively equates the findings of Doolittle's paper with respect to Factors IX and V. But Doolittle's paper actually says significantly different things about Factor IX and Factor V.

Regarding Factor IX, the paper tentatively concludes that it doesn't exist in the lamprey: "As in the case of the Trace database, no evidence was found in the draft assembly for a factor IX, despite exhaustive searching with factor IX sequences from a wide variety of species." Doolittle et al. then conclude "we would cautiously propose that a gene corresponding to factor IX is not present in the lamprey genome." (emphasis added) But like many of the other factors Miller tests for, Behe never claimed in Darwin's Black Box that Factor IX was part of the irreducibly complex core of the blood clotting cascade.

Regarding Factor V, the paper does not say that it is nowhere to be found in the lamprey genome. Miller claims that the lamprey "lacks" Factor V, but the paper says "the data are only consistent with a single copy of a gene corresponding to the precursor of clotting factors V and VIII." The paper thus doesn't say that the lamprey wholly "lacks" Factor V, but rather it says that the lamprey merely "lacks the separate equivalents of factors V and VIII (i.e., it has a preduplication gene)."

Behe did claim in Darwin's Black Box that both Factors VIII and V were part of the irreducibly complex core of the blood clotting cascade, so if the lamprey lacks one of those two factors, we might have a challenge to Behe's hypothesis. But even if the lamprey has only one gene that resembles both factors, it's difficult to conclude that one of the functions of the two factors is being unfulfilled. After all, many proponents of the gene-duplication hypothesis argue that first a gene acquires multiple functions, and then after a duplication event, the functions of the two now-separate genes diverge. So it's difficult for Miller or Doolittle to argue, under their theory's own terms, that at least one of the two Factor V/VIII's functions was going unfulfilled.

Indeed, Dr. Miller later admits in his response to me that according to Doolittle's paper, "The lamprey genome does contain a single gene, somewhat related to Factor X and Factor V, but not identical to either." (Miller here actually accidentally misstates the paper, which in fact states that "factors V and VIII seem to be represented by a single gene." If I made a mistake like this I'd probably be reamed for making some kind of "fail," but I'll gladly let Miller off the hook for his unintended mistake.) Miller's error aside, this is an important admission that could defeat his argument: taking Doolittle's conclusion at face value, the lamprey very well MAY still have a component that fulfills the functions of both Factors V and VIII.

But there is more...

Clearly, there IS at least one factor V-like protein in the lamprey genome -- and as discussed below, there is likely more than one. Searches for key domains from Factor V turned up dozens of "hits" in the lamprey genome. But Factors V and VIII are very similar, and under their view that there's only one such protein, the authors conclude that this one gene may represent a "precursor" to both proteins. An important question should now emerge:

Is there actually only one protein resembling Factors V and VIII in the lamprey genome?

If there is more than one protein resembling Factors V and VIII, then Doolittle and Miller's argument that the lamprey lacks one of those two proteins falls apart.

Sparse Data Backing Doolittle's Hopes and Dreams
Before I assess Doolittle and Miller's answer to the question I've just posed, we have to ask whether the question can even be adequately addressed given the current non-existence of a complete lamprey genome database. Doolittle is alleging that particular blood clotting factors do not exist in the lamprey genome, yet a continuous database of the entire lamprey genome has not even yet been created. That poses a problem.

To answer our question, Doolittle must work many fragments of lamprey DNA sequences in a "trace" database. The "trace" database does contain over 14 billion nucleotides of lamprey DNA, and the lamprey genome is on the order of 2 billion base pairs. Doolittle searched (and found) 20 lamprey genes sequenced apart from the "trace" database, and concludes that therefore the trace database is "sufficiently redundant to make judgments about the presence or absence of particular genes." Perhaps, but the bottom line is that the redundancy is, by Doolittle's estimate, only "fivefold," and we truly don't know for a fact that all lamprey DNA is in the database.

Despite the lack of a whole lamprey genomic database, Doolittle sets out to investigate some ambitious claims in his paper. Why? Doolittle's paper, seemingly, is one he has wanted to write for a long time. It is subtly framed as a response to Behe's argument for irreducible complexity. Doolittle even says that before the study, he "hoped" that the lamprey blood clotting cascade would be simpler than that of other jawed vertebrates because previously it was said that "every component seems essential":

Blood coagulation is known to follow a similar scheme in all vertebrates, the culminating event being the thrombin-catalyzed conversion of fibrinogen into fibrin. Interest in how the process evolved to yield the complex system that occurs in mammals has been longstanding, not only because every component seems essential, but also because some of the factors -- like factors IX and X -- depend on others -- like factors VIII and V -- for their activity. It has long been appreciated that a series of different gene duplications gave rise to many of the factors, and it was hoped that studies on early diverging vertebrates, especially jawless fish, might reveal a simpler process as it existed in earlier times.

(Russell F. Doolittle, Yong Jiang, Justin Nand, "Genomic Evidence for a Simpler Clotting Scheme in Jawless Vertebrates," Journal of Molecular Evolution, Vol. 66:185--196 (2008), emphases added)
At the very least, Doolittle is admitting that he entered this genomic study because he "hoped" that the lamprey might have a "simpler process" of blood clotting, thereby confirming his evolutionary bias. Such sparse data easily afford the opportunity to indulge your greatest hopes, and to confirm your biases. Looking at Figure 10 of Doolittle's paper, I wonder if he was disappointed to find a system that really is, at the end of the day, at most lacking only 2 elements of the jawed vertebrate pathway -- the V-->Va component, and the X-->Xa component. Studies that are undertaken after their authors "hoped" for a particular result, especially when the subject is data as sparse as is the case here, need to be carefully scrutinized.

Is there really only one protein resembling Factors V and VIII in the lamprey genome?
As noted above, if there is more than one protein resembling Factors V and VIII, then Doolittle and Miller's argument that the lamprey lacks one of those two proteins disintegrates. So let's now return to this important question: Does Doolittle's data support his claim that there is only one possible gene in the lamprey database resembling Factors V or VIII, and thus at most only one of the two proteins exists in the lamprey? I don't think it does at all. Why not?

Factors V and VIII are extremely similar, and Doolittle's paper admits that both factors yielded dozens of hits in the lamprey genome database. The proteins have a highly similar domain and sequence structure.

Both Factors V and VIII are composed of 3 types of components: three "A" domains (that are nearly identical), a "B" domain, and then a "discoidin" domain at the C-terminus. In both factors V and VIII, the A domains and "discoidin" domain are extremely similar, and the only major differences between the two factors lie in their "B" domains. But Doolittle's study searched and reported that dozens of hits similar to the A domains of Factors V and VIII were found in the lamprey genome! As the paper stated:

All told, searches of the six A domains (three each from human factors V and VIII) identified approximately 257 "hits" in the lamprey trace database. When these were compared with each other, many were found to contain the same (or virtually the same) inserts, and the number was subsequently reduced to 50 "hit-groups," consistent with an approximate fivefold redundancy. There was a wide range, however, the number of identical inserts in the various groups ranged from 0 to 29.

(Russell F. Doolittle, Yong Jiang, Justin Nand, "Genomic Evidence for a Simpler Clotting Scheme in Jawless Vertebrates," Journal of Molecular Evolution, Vol. 66:185--196 (2008), emphasis added, internal citations removed)

The known raw data thus actually shows that there is much more than merely ONE protein in the lamprey that resembles Factors V or VIII. The only reason that they don't tag many of these proteins as Factors V or Factor VII is because there are other proteins in the blood, namely hephaestin and ceruloplasmin, that also use these types of A-domains. Doolittle's reasoning, laden with evolutionary assumptions, ignores any protein that was allegedly more similar to hephaestin or ceruloplasmin as not possibly representing Factor V or Factor VIII, even though there were at least 50 such potential "hits" in the lamprey genome:
Of the 50 unique hits, fewer than a dozen resembled factors V or VIII more than or as much as they did hephaestin or ceruloplasmin. Concerned that the initial search might have been compromised by the long query sequences that are inconsistent with exon-length targets, we redid the entire exercise on an exon-by-exon basis, using human exon sequences as queries. ... An analysis of matching segments make it clear that, at a minimum, there are four genes in lamprey that belong to the hephaestin-ceruloplasmin-factor V/VIII family, only one of which seems closer to the FV/VIII side of the family."

(Russell F. Doolittle, Yong Jiang, Justin Nand, "Genomic Evidence for a Simpler Clotting Scheme in Jawless Vertebrates," Journal of Molecular Evolution, Vol. 66:185--196 (2008).)

Given that the paper reported "hits" for many other lamprey genes based upon finding matches for strings of a mere 30 amino acids (see Table 1), it would seem that the authors are applying more stringent criteria for detecting Factors V or VIII than they applied elsewhere. It's almost as if they "hoped" these factors wouldn't be found.

All those ignored "hits" aside, if one looks carefully at the data reported, it turns out that there actually wasn't just one protein that very closely resembled Factor V or Factor VIII, but TWO! The conclusion of their paper assigns a "hit" from the lamprey genome, Co-13587, to Factor V / VIII because it contains properties of those proteins including the unique "discoidin" domain at the C-terminus. But in Figure 4 of the paper, they note that another hit, Co-68932, "has its introns in exactly the same place as human factors V and VIII genes." Thus, according to their data, there would actually seem to be at least two proteins in the lamprey genome that are similar to Factors V and VIII.

Factors V and VIII comprise proteins, and the lamprey genome apparently contains two proteins that are highly similar to those to proteins. Coincidence?

I'm at a loss to see why Doolittle et al. claimed there is only one gene closely resembling Factors V/VIII when in fact there appear to be at the very least two, if not more. And if there appear to be two, then they have no reason to claim that one of those two factors doesn't exist in the lamprey genome. But of course, such a result might dash Doolittle's hopes and dreams.

A Deeper Core?
Though Doolittle's conclusion that the lamprey lacks an individual gene for Factor V appears to be unsupported by the data, let us now, just for the sake of argument, accept this conclusion. First, it shows that Behe's hypothesis of irreducible complexity is testable. But we have to test it realistically. While Doolittle's research is a good start, there could still be an irreducible core lurking at some lower level of complexity.

Miller seems more eager to refute Behe than he does to actually test the argument for irreducible complexity. Given that the lamprey still has nearly all of the components of the blood clotting cascade that Behe proposes are in the irreducibly complex core of the cascade, perhaps Doolittle's paper indicates that there is an irreducible core to the blood clotting cascade after all: The paper admits that in the lamprey genome, "The basic events involving the thrombin-catalyzed conversion of fibrinogen to fibrin and its subsequent cross-linking and lysis were not an issue." Indeed, Figure 10 has a proposed schematic of the lamprey blood clotting cascade that basically has all of the elements "after the fork" in the jawed vertebrate cascade, except for the V-->Va step and the X-->Xa step.

Importantly, the paper notes that "None of the principal clotting factors is found in the genome of the protochordate Ciona intestinales." So if the closest alleged relative to vertebrates -- the tunicate -- apparently has none of these principal clotting factors, could we be zeroing in on an irreducibly complex core? For those who take the ID hypothesis seriously, and aren't interested in false refutations of it, it's a possibility worth exploring.

But before the lamprey blood clotting cascade can be understood well enough to adequately address some of these questions, it seems we'll need more sequence data and a more realistic analysis of the lamprey genome than was employed in Doolittle's paper.

The lesson to be learned here is to always fact-check the claims of supremely confident defenders of Darwin like Dr. Ken Miller. He's a very smart biologist. However, sometimes looking closely at his citations shows just how weak his arguments are. In this case, it seems very likely that Miller's authority, Doolittle, had a very weak basis indeed for claiming that the lamprey lacked Factor V or Factor VIII.