12 November 2007

Peer review of my blog...let's see Behe try this.

Like every other scientist I know, I'm a big believer in peer review. The self-checking mechanism that peer review represents is surely one big reason for the success of science. Accountability, error checking, "wisdom in many counselors," and enforcement of community standards -- those are some ways of expressing the benefits of peer review. Some scientists, upon publishing their research, will thank the reviewers for making their article better. In his scathing review of Michael Behe's The Edge of Evolution in the New York Times, Richard Dawkins saved his most devastating criticism for last, by noting that Behe has

bypassed the peer-review procedure altogether, gone over the heads of the scientists he once aspired to number among his peers, and appealed directly to a public that — as he and his publisher know — is not qualified to rumble him.

--"Inferior Design," The New York Times, 1 July 2007

(Your irony meter should be pegged: Dawkins has famously done the same thing in The God Delusion. But that's another topic.) Peer review is a foundational principle in the scientific community, and those who eschew it are expressing outright contempt for the scientific enterprise.

Peer review isn't perfect, of course, and in fact it's only as good as the people who do it. It's not uncommon for us to get reviews that are wrong, even laughably so, or that have been written by "peers" who evidently didn't understand the work at all. And, on the flip side, the process isn't as blind or unbiased as it is often portrayed; scientists know how to slant their writing toward likely reviewers, and how to cherry-pick journals and potential reviewers in hopes of getting a better outcome. Yes, peer review is a human endeavor, with all the weaknesses of the humans doing it. But only a fool (or a demagogue) would consider doing science without it.

Well, here I am writing about science on a blog. One thing I really want is some peer review, at least to reduce the dangerously-high likelihood that I'll embarrass myself by posting something goofy. And so, I've been asking other scientists to read my articles. Specifically, after reviewing articles from the recent literature, I've contacted the authors and asked them to read the blog and provide comments. I've done four Journal Clubs so far, and the authors of two of them have graciously provided me with feedback. Now, these are top-flight scientists publishing in Science and Nature, and I was pleasantly surprised at their willingness to share some time with me. One of them is Joe Thornton, senior author of the two articles on steroid hormone receptor evolution that I recently summarized. With Joe's permission, I here present a summary of his review of my work.

One thing Joe didn't like at all was my gentle treatment of Michael Behe:

I think your description of Behe's argument is too generous. As you state, he doesn't argue that nothing can evolve in stepwise fashion, because selection can drive the evolution of such systems if each step increases fitness. But Behe does argue that integrated systems, in which the function of any part depends on the existence of the other parts, cannot evolve in such a fashion because selection cannot favor the origin, maintenance, or optimization of the parts until the entire whole is present. This argument is incorrect, because -- as we have shown -- such systems can be assembled by recruiting old molecules that previously had different functions to participate in new interactions, thus generating a new, integrated complex.

Joe's right about Behe's argument (with respect to irreducible complexity), and it's a lousy argument. But it is frequently misunderstood and oversimplified, and worse, right now, it is being erroneously conflated with the argument in The Edge of Evolution, which is actually different. My purpose in my seemingly too-polite comments about Behe's challenge was to direct readers to more serious engagement (and refutation) of Behe's claims and errors. Specifically, I wanted to draw attention to Larry Moran's work on Sandwalk, where he notes that some of the recent reviews of The Edge of Evolution have been grossly bungled, giving free shots to Behe and his attendant propaganda network. Dawkins, for example, in the NY Times review, is clearly aware of the mistake that Behe is making, but I think his piece is easily misunderstood (or twisted) to suggest that Behe doesn't believe in "microevolution." We can blame Behe for being unclear, even obfuscatory, and we'd be right, but that won't help us explain his damaging errors to non-scientists. For now, I'll risk seeming "too generous" to Michael Behe in order to ensure that I deal accurately and effectively with his carefully-packaged misinformation mistakes.

Joe did find some mistakes in my article, which I've now fixed:

  1. I claimed that his most recent work assembled a detailed family tree for the various steroid receptors. In fact, the 2006 paper presented and discussed that same tree.

  2. I claimed that the family tree was constructed from sequences of the two types of receptors, from 30 vertebrates. In fact, 29 species were involved, and the number of receptors known in each species ranged from just one to more than the two I was discussing. This was an error of simplification, and not very important, but it's been corrected.

And Joe noted, as I did, an oversimplification in the article:

It's an oversimplification to refer to the "corticosteroid receptor" and the "aldosterone receptor." For one thing, aldosterone is a corticosteroid. For another, the so-called aldosterone receptor exists in species that don't make aldosterone; in those species, it's generally a deoxycorticosterone (DOC) receptor -- another corticosteroid. DOC appears to be the oldest of the hormones and was probably the ligand for the ancestral receptor, before aldosterone itself evolved.

In this case, I've left the oversimplification as is, and encourage those who are irritated by it to read Joe's papers for the complete experience.

So there. My blog's been peer reviewed, to whatever extent Joe Thornton and I are "peers." :-) (I've made the cut on The Panda's Thumb twice, but unlike Joe, I've never been honored by the President of the United States for my work.)

But again with his permission, I'd like to share some of Joe's thoughts on the relationship between the "challenge" of ID and the work of real scientists like Joe (and me).

I'd like to be clear that, although the language we use to describe the question about the evolution of complexity may indeed be familiar to ID proponents, our work was in no way motivated by their arguments. This "puzzle" has motivated me since I began studying receptors, before Behe wrote his first book and before I had ever heard of ID. I continue to study the evolution of hormone-receptor evolution because it provides a superb system for unraveling the evolution of molecular complexity and for reconstructing the mechanisms by which gene functions evolved. The relevance of our findings to the social debate about ID didn't escape my notice, of course, and I didn't shy away from them; nevertheless, our research directions are motivated by evolutionary questions, not a desire to respond to ID.

This, I think, is one big risk entailed in the engagement of ID claims: that the magnificent science being done by Joe Thornton and hundreds of his colleagues would be portrayed as a "response" to ID. Good scientists, in my experience, tap into reserves composed mostly of intense curiosity, intellectual audacity (I consider that a compliment), and personal ambition. I'm horrified by the fact that ID's errors are linked to Christianity, and I'm willing to work on discrediting the movement, but I also know that this motivation could never fuel the kind of effort that generates science like Joe's or Chris Organ's or Abbie Smith's. My own work on cytoskeletal signalling systems in neurons could certainly be applied to ID claims in some way, but I'm not motivated by this at any discernible level. I just want to know how neurons work, and I want to be published more, and I want a renewal of my grant.

And finally, Joe noted the inverted perspective of ID with regard to explanation. When I talk to audiences about ID, I try to get them to consider this inversion by asking: "If 'intelligent design' is the answer, what was the question?" Design, it seems to me, is the thing we're trying to explain. It's not the explanation. Well, here's what Joe wrote in response to my blog entry:

You say, "And let's give ID credit for asking a good question." On one hand, I agree. Behe did, in parallel to us, identify the modern molecular version of the evolutionary puzzle of complexity: how can complex integrated systems, in which the function of any one part seems to depend on its interactions with the others, evolve under the influence of selection? Darwin was well aware of this puzzle, and the evolutionary geneticist H. Muller addressed it in a 1939 paper, explaining how complex systems that are historically assembled come to look "irreducible" as they evolve to be ever more functionally integrated. There have been many exemplars documenting the evolution of complexity at the morphological and physiological levels. But in recent decades, with the rise of molecular biology, innumerable systems comprised of tight functional interactions among molecules have been revealed. And few clear case studies were available to explain how complex systems evolve at the molecular level. So this is an important question indeed. But I am not convinced the ID proponents have ever really asked it. We reacted to this puzzle and knowledge gap as a question which stimulated a research program; after years of work, we now have some answers. The ID proponents, in contrast, put forth the issue of complexity as an answer, as an intrinsically unresolvable paradox that somehow gives the lie to Darwinian evolution. They do not ask, "How could complex systems evolve?" Rather, they argue, "Such systems cannot evolve by Darwinian mechanisms." And whenever scientists generate knowledge that begins to answer the question, the ID proponents bend over backwards to dismiss it. It appears this is a question to which they do not want an answer. That is a very big difference in the response to the apparent puzzle of complexity.

Very well said, don't you think?

So, there are some results of peer review of my blog. This week or next, I'll present the results of peer review in my blog. A couple of years ago, to much fanfare, Jonathan Wells published a paper on a topic I happen to know quite well. So I'll do some peer reviewing of my own, and we'll see whether ID really has produced something of scientific substance.

Please review my Rules and policies before posting a comment. Note that comments are closed after a month. If you would like to get in touch with me, visit the About page for contact details, including an anonymous comment form that works all the time.

blog comments powered by Disqus