A new wine delivery service called Vinome is promising to deliver “the ultimate personalized wine experience”—customized to your DNA.

There isn’t much (or, really, any) science to back it up. But it’s got a very big name in its corner.

Vinome just inked a deal with a startup called Helix, which in turn is backed by the world’s biggest DNA sequencing company, the powerhouse known as Illumina. For the past 15 years, Illumina has been selling machines that can quickly decode the human genome. Medical researchers around the world use them. But the company wants to conquer the consumer market, too. That’s why it spent $100 million to launch Helix, which teams up with app developers who can find creative ways to use a customer’s genetic data.

Such as selling them wine.

For about $65 per bottle, Vinome promises to pick out “great wines that are perfectly paired to you” based on an analysis of 10 genetic variants in your DNA, collected via saliva samples. The company—which is based, of course, in Northern California’s wine country—even incorporated the distinctive double helix of DNA into its logo of a corkscrew.

Medical geneticist Dr. Jim Evans isn’t impressed.

“It’s just completely silly. Their motto of ‘A little science and a lot of fun’ would be more accurately put as ‘No science and a lot of fun,'” said Evans, who’s a professor and researcher at the University of North Carolina.

“I’d put this in the same category as DNA matching to find your soulmate,” he said. “We just simply don’t know enough about the genetics of taste to do this on any accurate basis.”

Vinome is just one in a growing wave of targeted consumer genetic tests, which promise to deliver insights about nutrition, weight, and athletic training based on analysis of cheek cells, blood, or saliva. The tests could greatly expand the market for DNA analysis, to the benefit of companies like Illumina, which lately has been struggling to meet sales forecasts.

But public health experts see reason for concern. There’s rarely a straight line between a genetic variant and a complex trait like fitness, weight — or affinity for a good merlot. A complex interplay of genes and environmental factors shape our lives.

“Elevating genes to this kind of status for these kind of complex traits can be inappropriately distracting” from the important social and environmental causes of problems like obesity, said Tim Caulfield, a health policy professor at the University of Alberta who studies consumer genetic testing.

(On top of that, he said he saw a certain irony in adding a “deterministic vibe” to wine drinking. “It seems like the last place you’d want it,” he said. “I hope they don’t do that with beer.”)

‘Arguably defensible’ science

Vinome acknowledges the skepticism, but CEO Ronnie Andrews says he’s sure they’re on to something.

His team includes scientists with backgrounds in molecular diagnostics and cancer genomics.

“Listen, we wouldn’t risk our reputations to launch something in this space … if we didn’t feel like we had something that was arguably defensible in terms of the science, but more importantly would present an incredible experience and a fun experience,” said Andrews, who has formerly held executive roles at General Electric Healthcare and Roche.

The company soft-launched online in May and says it’s already shipped about 300 bottles of California wines to dozens of customers. (The minimum purchase is three bottles— plus genetic testing— for $199.)  The official launch will come this holiday season.

Next year, Vinome will roll out an app in partnership with Helix. In a statement provided to STAT, Dr. James Lu, Helix’s senior vice president of applied genomics, said his company “is excited to partner with Vinome because of their novel, fun and modern approach to building a wine recommendation algorithm.” Lu also compared Vinome to Netflix.

A promotional video on Vinome’s website features four people enjoying a picnic—with wine, of course — under a beautiful canopy of trees. “Most things are best when the guessing stops,” the narrator intones—just before a woman with light hair is matched with “aromatic, floral, bright whites” and a man with dark hair is matched with “powerful, textured, bold reds.”

Where do such recommendations come from?

Vinome surveys customers about their taste preferences and looks at their genetic variants in genes such as TAS2R38, a taste receptor. A small body of research suggests variations in markers on that gene can affect whether people can taste a bitter chemical found in certain vegetables—and as a result, whether they like Brussels sprouts.

Cofounder and lead scientist Sara Riordan acknowledges that her team found “virtually nothing” in the scientific literature that linked DNA variations with affinities for certain wines. Even the general literature on the genetics of taste and smell is “pretty small,” she said.

So Vinome decided to gather its own evidence. The team analyzed about 40 genetic variants across about 500 people. Then they had participants taste and rate a dozen wines and fill out a survey about their taste preferences.

The taste test was not designed with a control group—which is generally considered essential for reliable scientific evidence.

But all of it was enough to win over Helix. (Helix “understands that there are no genetic studies that show genetic polymorphisms related to individual wines,” Lu said. He also said Helix takes “the underlying scientific content in each of our products very seriously” and noted existing literature on taste genetics as well as the statistically significant associations found in Vinome’s taste test.)

Vinome has gone ahead and filed for patents and plans to try to publish its findings in a scientific journal, Andrews said. He’s not deterred by critics who point out that the science behind his venture is flimsy at best.

“I hear the skepticism,” Andrews said, “but those people just aren’t wine fans.”

Republished with permission from STAT. This article originally appeared on October 21, 2016