Lipskiy / Mad Dog / Shutterstock / The Atlantic

David Reif, now a biologist at NC State, realized his old paper had taken on a dangerous second life when he saw it cited—not in the scientific literature, but in a court case.

The paper was titled “Genetic Basis for Adverse Events after Smallpox Vaccination,” and it came up in 2016 when a vaccine-skeptical doctor tried to argue that it explained her patient’s development delays. The court was not persuaded, but Reif’s co-authors began hearing of yet other doctors using DNA tests to exempt patients from vaccines. Just this month, San Francisco’s city attorney subpoenaed a doctor accused of giving illegal medical exemptions from vaccination, based on “two 30-minute visits and a 23andMe DNA test.” On anti-vaccine blogs and websites, activists have been sharing step-by-step instructions for ordering 23andMe tests, downloading the raw data, and using a third-party app to analyze a gene called MTHFR. Certain MTHFR mutations, they believe, predispose kids to bad reactions to vaccines, possibly even leading to autism—a fear unsupported by science.

This interest in MTHFR can be traced right back to Reif’s 2008 paper, which linked a variant of the gene to “adverse events” after smallpox vaccines. It was a somewhat intriguing result then. A decade later, however, James Crowe, the senior author of the paper and the director of the Vanderbilt Vaccine Center, offers a blistering assessment of his own study: “It’s just not even a valid study by today’s methodology.” To use it for granting vaccine exemptions now, he says, is “illogical and inappropriate.”

What’s changed? The basics of genetics research. With MTHFR, opponents of vaccines have been able to exploit a lag between the advance in scientific knowledge and widespread understanding of it. In the past 15 years, mainstream genetics research has evolved—but in ways that are not always obvious to the public.

A lot of that confusion is in fact centered around MTHFR. In the early days of genetics research, scientists looking at a small number of genes in a small number of people found that certain MTHFR variants were linked to a range of maladies: blood clots, cancer, heart disease, pregnancy complications. This seemed to make sense, as MTHFR codes for an important enzyme in the body. But as scientists went from looking at hundreds to thousands to hundreds of thousands of people, they realized many of those variants were extremely common, found in up to 40 percent of the population in some cases.

More importantly, those associations with various diseases just didn’t hold up in bigger data sets and with better statistical tools. They had been flukes all along. MTHFR is not alone in this. The scientific literature is littered with “candidate genes” that turn out not to explain much at all. The American College of Medical Genetics and Genomics currently does not recommend testing for MTHFR.

But the word about MTHFR was already out. And the bold claims about the gene have lingered on alternative and naturopathic medicine sites. “It’s been hard to reverse because if you go to historical literature, you can find some literature to support your claim,” says Elizabeth Varga, a genetics counselor at Nationwide Children’s Hospital. “Without a big picture, without knowing the full story, I think that’s where people who want to exploit the information, they can exploit the information.” And now, mail-in DNA tests have made it relatively easy for anyone to get tested.

In 2017, 23andMe published a blog post noting that MTHFR was the “most asked-about gene by 23andMe customers.” Its scientists concluded that common MTHFR variants were not clinically useful to test. (That’s why anti-vaccine doctors have to download the raw 23andMe data and analyze MTHFR using a third-party tool such as Genetic Genie or Promethease.) The company told me it does not condone the use of its test to grant vaccine exemptions and pointed toward a disclaimer about using its raw data for medical or diagnostic reasons.

Interest in MTHFR and vaccines specifically seemed to have ticked up in 2016, when California outlawed personal and religious exemptions from vaccination. Doctors could still grant medical exemptions, however, and anti-vaccine websites started sharing possible reasons for doing so. MTHFR was listed as one of several, and for evidence they linked to Crowe and Reif’s paper. (Interestingly, the paper actually identified a second gene called IRF1 as well, but the focus has always been on MTHFR, perhaps because the latter gene was already familiar to the naturopathic community.)

Crowe says that understanding how genetics affect reaction to vaccines is still a legitimate area of research, but the 2008 paper was emblematic of that early, now outdated genetics research. The paper was composed of two small studies, one with just 85 participants and the other with 46. To get published today, such studies would likely need thousands of participants and to have validated the results in a second group of people. “We were just starting to figure out how to use the genome,” Reif says. The tools they used—along with so many other geneticists at the time—were simply not up to snuff.

Furthermore, the study only covered smallpox vaccines, which are no longer given to children, as the disease has been officially eradicated since 1980. It wouldn’t make sense to extrapolate to measles vaccines given today. And the “adverse events” were simply mild fevers and rashes—nothing that would bolster the discredited link between vaccines and autism.

That vaccine-autism link also infamously originated with a single journal article, now retracted. The anti-vaccine movement is decidedly outside of mainstream medicine, but it has always borrowed the language and trappings of mainstream science. By tapping into the wider interest in genetics, vaccine skeptics are attempting to tap into scientific legitimacy. The early hype about the power of genes and the early spate of now outdated research made genetics research all the more exploitable.

Plenty of solid genetics research has been published since then, but the idea that genes are powerful has spawned a world of less rigorous ideas about DNA too. At a time when companies are hawking (unproven) DNA-based diets, supplements, and exercise routines, of course people are wondering what genetics says about vaccines—even if the answer is, currently, not very much at all.

Crowe and several of his co-authors are now drafting a letter to the journal that published the 2008 paper, clarifying just how little their original paper actually says. And on Wednesday, the California Senate advanced a new vaccine bill making it harder to get medical exemptions. Public-health officials—not individual doctors—would have the power to grant them.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.