The Fact-Free "Science" of Matheson, Hunt and Moran: Ridicule Instead of Reason, Authority Instead of Evidence
I was not in Los Angeles on May 14, when Stephen Meyer debated Stephen Matheson and Arthur Hunt at Biola University. But I have followed some of the blog war that preceded and followed the debate--a blog war that now includes Richard Sternberg and Laurence Moran.
Since Matheson, Hunt and Moran are all tenured professors at institutions of higher learning, one might have expected a discussion based on reason and conducted in a collegial spirit. And since the discussion is about science, one might have expected lots of references to evidence published in the scientific literature. But Matheson, Hunt and Moran have abandoned reason and resorted to ridicule; and instead of citing evidence they expect us to bow to their authority.
Round One: Who's Who
Meyer is a philosopher of science and the director of the Discovery Institute's Center for Science and Culture (with which I am also affiliated). Meyer wrote Signature in the Cell, which was the focus of the May 14 debate. Matheson is a neuroscientist at Calvin College who has been doing research on formin, a protein affecting the network of microscopic fibers that give a cell its shape. Hunt is a plant biologist at the University of Kentucky who does research on the processing of messenger RNAs, the molecular intermediates between DNA and protein.
Before the debate, in a February 14 blog post, Matheson accused Meyer of "some combination of ignorance, sloth, and duplicity" for stating that introns, though once thought to be "junk," are "now known to play many important functional roles in the cell." (Introns are non-protein-coding segments of DNA that interrupt the protein-coding parts of our genes.) According to Matheson, Meyer's statement was "ludicrous" because biologists have identified functions for only a "handful" of the 190,000 or so introns in the human genome. "How many? Oh, probably a dozen," wrote Matheson, "but let's be really generous. Let's say that a hundred introns in the human genome are known to have 'important functional roles.' Oh fine, let's make it a thousand."
In a June 3 blog post, Richard Sternberg responded to Matheson. Sternberg is an evolutionary biologist and systems biologist (he holds a Ph.D. in each field) at the Biologic Institute (with which I am also affiliated). From 2001 to 2007, Sternberg served as a staff scientist at the National Center for Biotechnology Information and a Research Associate at the Smithsonian's National Museum of Natural History. He has published scientific articles with James Shapiro, a molecular biologist at the University of Chicago, on the functions of repetitive DNA, and he recently published a scientific article on the relationship between DNA and biological information.
In his June 3 blog post, Sternberg cited a 2010 Nature article that begins, "Transcripts from approximately 95% of multi-exon human genes are spliced in more than one way." (Exons are the protein-coding parts of genes.) Sternberg calculated that a reasonable estimate for the number of human introns that undergo alternative splicing was thus 0.9 x 190,000 = 171,000. The same Nature article reports that "in most cases the resulting transcripts are variably expressed between different cell and tissue types," suggesting that many alternatively spliced introns serve important biological functions. Sternberg concluded that even if his estimate were off by a factor of two, the number of functional introns would still be far greater than Matheson's 1,000.
At that point the debate spilled over into Laurence Moran's blog, Sandwalk. Moran is a biochemist at the University of Toronto who published research in the 1990s on proteins whose expression is increased when cells are exposed to elevated temperatures. Moran now co-authors textbooks on biochemistry and writes on the Internet about molecular evolution and evolutionary theory.
Round Two: Insults and Ridicule
In a June 3 blog post, Moran ridiculed Sternberg's estimate that 171,000 introns are involved in alternative splicing: "It's up to you, dear readers, to figure out all the things wrong with this explanation. You can start with the math. Arithmetic isn't one of their [i.e., Meyer's and Sternberg's] strong points. Or maybe it's an understanding of biology that's the real weak point?"
Of course, there was nothing wrong with Sternberg's arithmetic, and the next day he repeated his calculation in more detail. Moran responded by calling him an "idiot." Hunt and Matheson joined the fray by posting comments critical of Sternberg on Moran's blog.
Since this is a scientific debate, one might expect to see lots of references to evidence published in the scientific literature. Maybe I missed something, but I found only one reference cited by Matheson, Hunt and Moran. On February 14, Matheson cited a 2005 article in Current Genomics in support of his statement that "the human genome contains at least 190,000 introns." On June 3 (as we saw above), Sternberg used Matheson's figure of 190,000 as the starting point for his argument that a substantial number of human introns serve important biological functions. Sternberg then cited six other scientific articles in support of his argument. Except for the 2005 article on which Sternberg based his calculation, however, Matheson, Hunt and Moran have cited no scientific literature in this debate.
Since the debate also turns on how one defines "function" in DNA, one might expect to see a calm and reasoned discussion of that concept. As we shall see below, Matheson and Hunt touched on this issue, but they misrepresented and ridiculed Sternberg's position.
Round Three: Who You Gonna Believe?
In his June 4 blog post calling Sternberg an idiot, Moran wrote:
I'm using arithmetic that's based on an understanding of basic molecular biology and the scientific literature. We could quibble about the number of introns--I think it's closer to 150,000. We could quibble about the number of protein encoding genes--the most accurate number is 20,500. We could quibble about how many genes exhibit alternative splicing--I think it's about 5%, not 95%. You can't be expected to know the facts and the controversies since this is way outside your area of expertise.
So, let's assume your facts are correct. If 90% of genes exhibit alternative splicing then this means 22,500 genes. You got that calculation right. The minimum number of introns that must be involved in alternative splicing is one (1) per gene. That means at least 22,500 introns involved in alternative splicing. You made the mistaken (and stupid) assumption that every intron in a gene had to be alternatively spliced.
Now, 22,500 is a lot more than the 1,000 begrudgingly allowed by Matheson. Of course, some introns involved in alternative splicing could be just inert spacers--as Sternberg acknowledged on June 3. But the 2010 Nature article cited by Sternberg also provided evidence that introns are "rich in splicing-factor recognition sites," and the authors of that article predicted "regulatory elements that are deeper into introns than previously appreciated." So even if only 22,500 genes were alternatively spliced, and splicing involved on average only one intron per gene (not true, as we shall see below), the fact that many introns contain splicing-recognition sites suggests that the number of functional introns would be much greater than Matheson's 1,000.
Moran claimed that he was basing his remarks "on an understanding of basic molecular biology and the scientific literature," but he didn't cite any scientific literature. He simply declared that the number of human introns is closer to 150,000--even though the scientific literature supports Matheson's figure of 190,000.
And Moran's guess that only about 5% of human genes undergo alternative splicing is flatly contradicted by 2008 articles in Nature and Nature Genetics, as well as the 2010 Nature article cited by Sternberg. In lieu of factual support for his claim, Moran ridiculed Sternberg for giving "no indication that he understands the controversy" and basing "his entire fairy tale on a value [i.e., 95%] that has been pretty much discredited."
Yet Moran provided no justification for his ex cathedra pronouncement that the 95% figure has been discredited. He simply brushed aside the 2008 and 2010 articles in Nature and Nature Genetics and their eighteen co-authors--nine of whom listed their affiliation as Moran's own institution, the University of Toronto.
Apparently, Moran expects us to ignore the published scientific evidence and bow to his authority alone. As Chico Marx said in the 1933 movie Duck Soup: "Who you gonna believe, me or your own eyes?"
Round Four: Where's the Beef?
On June 3, Hunt posted comments critical of Sternberg on Moran's blog. According to Hunt, Sternberg was wrong for assuming in his calculations that "every intron in every gene that gives alternatively-spliced mRNAs will be subject to alternative processing. That's not very likely." Indeed, Hunt insulted Sternberg by accusing him of making "errors that I expect my students to avoid."
Yet Hunt cited no scientific literature, just as Moran cited none to support his assumption that only one intron per gene may be involved in alternative splicing.
Actually, a 2008 article in Science reported 94,241 splice junctions in human messenger RNAs. Since each splice requires two junctions, and humans probably have between 20,000 and 25,000 genes, this suggests an average of 2 introns per gene involved in alternative splicing. In 2009, scientists reported that the majority of human genes generate almost eight different messenger RNAs. But in order for a gene to generate eight alternatively spliced messenger RNAs it must contain at least three protein-coding segments, and thus at least two introns.
So the published scientific evidence indicates that alternatively spliced genes contain at least two introns--not the one intron assumed by Moran. This would yield a lower figure than Sternberg's original estimate--45,000, to be exact--but this is still a lot more than 1,000.
Yet two introns per gene is surely a lower bound. A chicken gene involved in hearing generates more than 500 alternatively spliced messenger RNAs, and a fruit fly gene involved in cell adhesion can potentially generate more than 38,000. In these cases (and others), more than 2 introns would be needed.
Any way you look at it, Matheson's estimate of 1,000 is far too low.
Round Five: What's in a Word?
What does it mean to say that an intron is "functional"?
In June 4 comments on Moran's blog, Hunt wrote that since intron splicing codes tend to be close to protein-coding regions, it is the latter that are functional, while the other parts of introns are "junk." (Of course, whether the other parts are "junk" is one of the matters under debate. As we shall see below, there is evidence against it.) Hunt then accused Sternberg of creating a "caricature of the alternatively-spliced intron as a unit that teems, end-to-end, with functions." Of course, Sternberg never claimed anything of the sort. The caricature is Hunt's.
In a June 5 blog post, Matheson acknowledged that some introns "do harbor functional elements, and perhaps many do, and while much certainly depends on what we all mean when we assign 'functional roles' to large chunks of non-coding DNA, it's only fair to acknowledge that [Meyer's] statement is not strictly false."
Matheson also wrote that Sternberg is "right that alternative splicing is an important and regulated process, and that intron sequences are involved in that process, and if that's all it takes for an intron (which can be far bigger than the splice junctions themselves) to be called 'functional,' then there are a lot of 'functional' introns in the human genome. But of course that's not what I and others mean when we question whether introns typically have 'function'." When he and Moran "assert that introns don't tend to have 'functions,' we mean that the huge amounts of sequence within the introns don't tend to have 'functions'."
But splice junctions are not the only functional parts of introns, as Sternberg pointed out in his first blog post on June 3. Introns encode microRNAs that are not translated into proteins but help to regulate specific messenger RNAs. Introns also generate non-coding RNAs that affect gene expression by modulating chromatin, the combination of DNA, RNA and protein that makes up a chromosome. And more functions of introns are being discovered all the time.
Round Six: Chest-Thumping in Lieu of Evidence
So why are Matheson and Moran so sure that huge portions of introns don't have functions? According to Matheson, it's because "Larry Moran and I clearly know a whole lot more about molecular genetics" than Sternberg.
A more naked appeal to authority would be hard to find. It sounds like an undergraduate trying to score points in a late-night bull session ("I know all about that; I took a course in it..."), not a college professor engaged in a scientific debate.
But Matheson didn't stop there. He demeaned Sternberg by calling him "poor Richard." He also claimed that Sternberg is "disastrously clueless" because he doesn't understand "the important and very basic distinction between a transcript and an intron." Since every undergraduate biology student learns that an intron is a segment of DNA, while a transcript is a segment of RNA encoded by DNA, this last jibe is on a par with Moran's insult that Sternberg can't do elementary arithmetic. And it is equally unjustified.
If one overlooks the nastiness, it is clear that there are some interesting issues in this debate. Conceptually, what does it mean to say that a segment of DNA has function? Empirically, what does the evidence show?
One might think that professors Matheson, Hunt and Moran would address the conceptual issue calmly, rationally, and collegially. But they don't; instead, they stoop to misrepresentation and ridicule. And one might think that they would address the empirical issue by citing published scientific evidence. But they don't; instead, they simply proclaim themselves the only authorities on the subject.
Who you gonna believe, them or your own eyes?