The now-famous 2009 Science paper by Lombardi and colleagues  showing a strong correlation between the human retrovirus XMRV and chronic fatigue syndrome (CFS) did not parachute into the middle of nowhere. As Hilary Johnson’s Osler’s Web recounts, feelings have been running high on this subject since the 1980s. Careers have been devoted—especially, but not only, in the United Kingdom—to the idea that CFS is not a physical disorder but a psychological one. The Science paper was bound to be unpleasant reading for anyone who had treated scores of CFS patients on the psychological theory and put their professional credibility on the line to defend that theory.
Opponents of physical-cause theories of CFS were therefore cheered by the appearance in January, 2010 of an article appearing to refute the Science piece. The new article, “Failure to Detect the Novel Retrovirus XMRV in Chronic Fatigue Syndrome,” by Erlwein et al.,  was hailed as showing that the Science piece was a false alarm: “Scientists’ claim to have found the cause of [CFS] is ‘premature’,” headlined The Independent, a British newspaper.  Nothing to see here, folks—it’s all over—move along, move along.
But as the Cylons repeat so annoyingly on Battlestar Galactica, “All this has happened before; all this will happen again.” Indeed it has, and indeed it will. Consider a recent case involving another highly charged subject, genetic engineering of crops.
Act I: Threatening Breakthrough
In 2001, David Quist and Ignacio Chapela of U.C. Berkeley announced in Nature that transgenes—chunks of DNA artificially transferred from one species to another—had been found in traditional maize landraces in central Mexico.  (A “landrace” is any locally-adapted domestic plant breed.) The study seemed to show that modified genes could spread uncontrollably in the real world, as opponents of genetic engineering had always warned. If so, transgenes might threaten the character or continuance of the Mexican maize landraces, altering the Mexican diet and the global fate of corn itself. As Quist and Chapela put it,
Concerns have been raised about the potential effects of transgenic introductions on the genetic diversity of crop landraces and wild relatives in areas of crop origin and diversification, as this diversity is considered essential for global food security. Direct effects on non-target species and the possibility of unintentionally transferring traits of ecological relevance onto landraces and wild relatives have also been sources of concern.Act II: Triumphant Counter-Study
Genetically modified corn is big business: 25% of the world’s corn, including 80% of US corn, is genetically modified.  It is therefore not surprising that the 2001 paper was intensely attacked for minor methodological flaws. Under pressure, Nature took the unprecedented step of publishing a quasi-retraction (not approved by the authors, Quist and Chapela) stating that “the evidence available is not sufficient to justify the publication of the original paper.” 
The debunking process seemed complete in 2005, when a paper in the Proceedings of the National Academy of Sciences reported finding not a single transgene in thousands of maize samples from the same parts of Mexico examined by Quist and Chapela: zero, repeat, zero transgenes in 153,746 sampled seeds. 
In certain quarters, the relief was palpable. The zero result was taken as grounds not for questioning the second study but for scolding the first. In a 2006 review  of the preceding decade’s nine worst “biotech gaffes,” the editor of Nature Biotechnology spanked Quist and Chapela for being partly “culpable for lukewarm public acceptance and stigmatization of transgenic crop technology.” That’s right, culpable as in deserving of blame. “With so many of the groups ideologically opposed to transgenic crops able to exploit the media, scare the public and perpetuate myths and conspiracy theories about genetic engineering over the Internet,” the editor harrumphed, “prestigious journals [e.g., Nature] should be aware of the long-lasting damage resulting from their willingness to widely publicize results that may be contentious or equivocal.” The maize-transgene scare had “certainly contributed to the decline of European agbiotech.” Very, very naughty.
In the backlash against the original article the academic career of one of its authors, Chapela, was almost destroyed by denial of tenure, but he won in court. 
Act III: Vindication
In late 2008, a paper in the journal Molecular Ecology  vindicated Quist and Chapela on the basis of tens of thousands of Mexican maize samples. Once again, transgenes had been found in Mexican corn landraces. What is more, the authors of the new study explained exactly why different studies had been getting different answers and explained the false-negative results of the 2005 Proceedings of the National Academy of Sciences paper in detail. Even the lead author of that 2005 paper, invited to comment in the same issue of the journal, sportingly pronounced the new piece “a very good study.”
It does not appear, however, that the editor of Nature Biotechnology ever retracted his attacks on Quist and Chapela’s integrity. Indeed, I know of no apology or retraction from anyone who had accused Quist and Chapela of bad science.
The parallels of the Mexican transgene saga to the XMRV saga are striking: In Act I upsetting primary science appears and solemn warnings against taking a single study too seriously are issued. In Act II, a single study offering to overturn the original is hailed with relief and trumpets. In Act III, the primary science is vindicated—at least, it was with the maize transgenes. There has been no vindicating third act yet for the XMRV/CFS theory, but there are at least three reasons to bet on one.
First, getting a zero result where multiple, independent previous studies have got a nonzero result should be a red flag with any new XMRV study, just as it should have been during the Great Mexican Maize Mystery. Zero? Are you sure you’ve taken the lens cap off the camera?
Second, both zero-result studies—the maize study and the UK XMRV study—used tests for the presence of the target gene or virus different from those used in the studies they challenged. In the case of the maize transgenes, this turned out to be the crux of the problem. Whittemore Peterson states that “the recent study published in the U.K. . . . used non-validated PCR and whole blood PCR assays.”  Moreover, the UK study did not select candidates for study using the Science study’s standard, thus producing a twofold apples-and-oranges problem.
Third, the UK XMRV study was rushed through peer review in a few days, according to the Whittemore Peterson Institute.  [PS: Actually, as a friendly commenter points out, this is according to the online journal where the piece was published: here’s a screen shot from the source:]
Science is done by human beings, not gods or robots. Its glory is that its method of community-scale, independent checking and criticism almost always enables the production of increasingly accurate knowledge, over time, by a group of people—scientists—who are not increasingly wise or perfect. But in the short term, especially where powerful economic and other interests are involved, the data that some people want to see have a tendency to appear—or the data they do not want to see may be long delayed by diversion of funding and other tactics. Reality wins in the end, but the end may be a while in coming.
Extraordinary claims do require extraordinary evidence, and it is easy to be jerked around by a single paper here and a single paper there. But in the case of the XMRV imbroglio, my money is on the Whittemore Peterson people, who have clearly done the more thorough work. In any case, more studies are under way, and if they are of adequate quality, they will settle this dispute—scientifically.
Unfortunately, that will not necessarily settle the dispute altogether. To answer any scientific question in a way that some people strongly dislike seems to automatically give birth to a new form of denialism. Evolution and global warming already have millions of entrenched, unconvincable unbelievers: is XMRV next?
Note: I will e-mail copies of the subscription-required articles to anyone who asks me (Larry) for them: lnpgilman [ a t ] wildblue [ d o t ] net
4. Quist, David and Ignacio H. Chapela, “Transgenic DNA introgressed into traditional maize landraces in Oaxaca, Mexico.” Nature, 414 (Nov. 29, 2001), 541–543.
7. Ortiz-Garcia, S., et al., “Absence of detectable transgenes in local landraces of maize in Oaxaca, Mexico (2003–2004).” Proceedings of the National Academy of Sciences, 102(33), August 30, 2005, 12338–12343. Available at http://www.pnas.org/content/102/35/12338.full.pdf.
10. Piñero-Nelson A., et al., “Transgenes in Mexican maize: molecular evidence and methodological considerations for GMO detection in landrace populations,” Molecular Ecology (2009) 18, 750–761.