07 March 2008

Hugh Ross' shocking fairy tale

“I was to some degree trusting that the vetting process of a reputable book publisher was going to catch this level of duplicity.” But, he added: “Do I wish in retrospect that we had called L.A. child services and tried to run down the history of this person? I certainly do.”
That's Tom de Kay, editor of the Home & Garden section of the New York Times. Last week Thursday, that section ran a story, "A Refugee from Gangland," describing the life of Margaret B. Jones, the author of a just-released "heart-wrenching memoir" set in gangland L.A. The Times piece is fascinating, and the memoir probably is too. One little problem: the memoir has just been revealed to be a fraud. It was wholly fabricated.

In the aftermath, editors and publishers and even journalists are asking hard questions, and the book is being recalled.

It's surprising, jarring, in many ways incomprehensible, and it's just the most recent example of a "gritty memoir" that turned out to be a slick work of fiction.

Some folk science is truly fiction, but it's not that often that one uncovers a cynically fabricated bit of history. And maybe I'm too much of a moral relativist, but I do see a difference between, say, selective citation of the scientific literature in support of a weak or false proposition and, say, completely inventing a story of scientific discovery that paints one's opponents as fools and one's colleagues as brilliant heroes. Let's see if you agree.

IMPORTANT NOTE: this post, seeking to be harshly critical of Hugh Ross, refers to some of his statements as "lies." Please read the rest of this post in conjunction with a more recent post, "On folk science and lies: back to the basics." There I respond to some very important criticism, and agree that "lie" is not a useful or appropriate term here.

In Creation as Science (NavPress, 2006, pages 168-168), Hugh Ross relates a story of scientific discovery that is nothing other than a slick work of fiction. Here's how it begins:
The assumption that the non-protein-coding part of the genome served no purpose caused researchers to abandon study of its features for nearly three decades. Then a team of physicists made an observation that revived interest. They noticed that the quantity of "junk" in a species' genome was proportional to that species' degree of advancement.
You already know that the first sentence is a common falsehood. But here's an interesting new twist: a "team of physicists" somehow "revived interest" in the study of non-coding DNA. Ross claims that they "noticed" a proportional relationship between "junk DNA" and "degree of advancement."

The research to which Ross refers is reported in this brief paper: R.N. Mantegna et al., "Linguistic Features of Noncoding DNA Sequences," Physical Review Letters 73:3169-3172, 1994.

First, the authors of the article in question represent two departments: the physics department at Boston University, and the cardiovascular division of Beth Israel Hospital and Harvard Medical School. The group is truly interdisciplinary, evenly split between the two departments, but Ross elects to refer to it as a "team of physicists," and I think that says a lot about what might explain his egregiously error-filled forays into biology. In fact, one of the coauthors (Ary L. Goldberger) is quite well-known as a cardiologist and the director of the Margret & H. A. Rey Institute for Nonlinear Dynamics in Medicine at Beth Israel Deaconess and Harvard Medical School.

We'll soon see how the claim that this paper "revived interest" in "junk DNA" is an outrageous lie, but what about the proportional relationship between non-coding DNA and "advancement" that Ross associates with the authors of the 1994 paper? Well, it's a very short paper, but here's the only sentence that Ross could possibly have in mind:
An intriguing puzzle is related to the fact that in higher organisms, only a small fraction of the DNA sequence is used for coding proteins; the possible function – if any – of the noncoding regions remains unclear [5].
(The reference there, by the way, is to a review article on "Introns as Mobile Genetic Elements." It's a 37-page survey of one particular class of non-coding DNA, published more than a year before the "team of physicists" was able to "revive interest" in the "study of its features." This should make you laugh, but it's really not that funny.)

Note that the authors of the 1994 paper did not claim that the "quantity of 'junk' in a species' genome was proportional to that species' degree of advancement." Only an ignoramus would have written that, because it's not true; indeed, it's so far from the truth that precisely the opposite is actually the case.

"Degree of advancement" is not a phrase a biologist would use, but let's assume Ross means "complexity," in the way that a giraffe is more "advanced" than a worm, which is more "advanced" than an onion, which is more "advanced" than a mushroom, which is more "advanced" than a bacterium. For decades, biologists have known that the amount of DNA in an organism is utterly unrelated to its complexity. In fact, the very notion of "junk DNA" (referring specifically to pseudogenes, at the time) was invented as a hypothesis to account for the surprising lack of any correlation at all between an organims's DNA content and its size or complexity (or, for that matter, its phylogenetic position relative to other organisms). This observation was so surprising in its time that it was termed a paradox: the C-value paradox.

In other words, Hugh Ross begins his little story with a statement, falsely attributed to an interdisciplinary research group that he inaccurately calls a "team of physicists," that is so stunningly far from the truth that it is incomprehensible as anything other than an outright fabrication. I don't see how it could be a mistake, but perhaps when Reasons To Believe starts issuing retractions and apologies for its myriad falsehoods, Ross will attempt an explanation. (For the record, I'd settle for a correction, an apology, and a pledge to uphold at least minimal standards of academic integrity.)

Believe it or not, it gets worse. Here's the rest of Ross' fabricated fable:
The physicists decided to perform a computer analysis, and in 1994 they published their results. They found that what had long been labeled junk DNA carries the same complex patterns of communication found in human speech. In fact, they found that the junk DNA had an even higher linguistic complexity than did the protein-coding DNA. This breakthrough discovery drew teams of geneticists worldwide into a veritable frenzy to uncover the hidden designs and functions of the portion of DNA once thought useless.

This flurry of research has revealed five kinds of noncoding (for proteins) DNA, and each kind plays an important role in the vitality and function of the organisms in which they reside...
"Breakthrough discovery?" Well, according to Google Scholar, that paper has been cited 187 times, and when I examined this using Scopus, I found that about half of the paper's citations are from biology journals. Most of the remaining citations are from journals focused on physics, computation, and information theory, and most of those are mainly interested in the computational aspects of the 1994 study, not in its implications for genomics or genetics. According to Ryan Gregory, an actual expert in the area of genomic evolution and genome size, the report had no discernible impact on the study of genomes:
It would seem that other computer and physics types were interested, but few mainstream genetics authors picked up on it. Some people challenged it as being an artifact (Bonhoeffer et al. 1996a,b), but mostly I think people dismissed it as wishful thinking, if they even heard of it.
– Prof. T. Ryan Gregory, interviewed by email
And in case you're wondering whether 187 citations since 1994 is a lot, consider that my most cited paper, a 2001 article on p190RhoGAP on which I am co-first author with Madeleine Brouns, has been cited 108 times since then, all by biology journals. It's a very good paper, and it reported some very important results, but I don't think any of my colleagues would call the findings "breakthrough discoveries." What would a 1993-4 "breakthrough discovery" look like? Well, remember microsatellites? They comprise just one interesting class of non-coding DNA that was being intensively studied during the 30 years that Ross claims were lost to science. In 1993, a group from the Mayo Clinic in Minnesota reported that changes in microsatellite DNA are a frequent occurrence in colon cancer. Their paper has been cited almost 1500 times since then. And that review article on introns, which is cited by the "team of physicists" above? It's been cited 324 times since then.

It wasn't a breakthrough; it wasn't even an important or particularly useful result. But the most shocking and disturbing aspect of Ross' fairy tale is the full-blown lie at the end. Hugh Ross claims, falsely, that this insignificant little article launched a "veritable frenzy" of research "worldwide" on the functions of non-coding DNA. A quick look at the trajectory of research in any area of genomics or molecular genetics reveals this to be laughably untrue, but the final proof that Hugh Ross needs to re-examine his basic integrity is that last sentence. He actually claims, in print, that the "veritable frenzy" of research unleashed by the "team of physicists" led to the discovery of the various classes of non-coding DNA:
This flurry of research has revealed five kinds of noncoding (for proteins) DNA, and each kind plays an important role in the vitality and function of the organisms in which they reside...
Now, Ross' list is (you guessed it) not accurate; he doesn't even mention introns, for example. But the jaw-dropping lie, of course, is the claim that the imaginary impact of the "team of physicists" led to research that "revealed" these non-coding DNA elements. I'll leave you with a list of Ross' five classes of non-coding DNA, and references to the reports of their discovery.

Pseudogenes
SINEs
LINEs
Endogenous retroviruses
LTRs

All of those non-coding elements were discovered more than 10 years before the little paper by the "team of physicists." (UPDATE: Ryan Gregory informs me that some of my dates are too generous: pseudogenes were known by 1977, and Alu elements (SINEs) were described by 1979.)

It's sad but true: you can't believe Reasons To Believe.

And if you work there, and you're reading this blog right now, please do something about this – for Christ's sake, if not for that of your own dignity.

blog comments powered by Disqus