Illustration: Wenkai Mao for Bloomberg Businessweek

About three years ago, Orig3n Inc., a small genetic testing company in Boston, discovered a big problem with its DNA analysis, according to two former managers, a former lab technician, and another former employee. Its various tests ($29 to $298) are supposed to help you figure out which kinds of food, exercise, and beauty products are right for your particular genetic profile. They’re even supposed to identify your “superhero traits,” meaning whether you’re genetically predisposed toward, say, intelligence or strength. The problem was that its test results were prone to errors. A person taking the same test twice could get radically different results.

Rather than try to solve the underlying problem, Orig3n’s coders came up with a quick fix, according to the former workers. The separate nutrition and fitness tests, for example, analyzed some of the same genes. The former workers say that if two Orig3n analyses of a particular gene didn’t match, software plugged in the earlier result. A spreadsheet viewed by Bloomberg Businessweek shows 407 such errors that, according to the former lab tech, were logged over three months.

Seventeen people employed by Orig3n as managers, lab technicians, software engineers, marketers, and salespeople from summer 2015 to fall 2018 say the company habitually cut corners, tampered with or fabricated results, and failed to meet basic scientific standards. They say marketers tasked with giving consumers personalized advice based on their genetic profiles at times simply Googled advice that could be generic (wear sunscreen, eat kale) or had no basis in science (one report told a customer to eat a mix of sugar and almond oil to reduce stretch marks). “Accurate science didn’t seem to be a priority,” says the former lab tech, who, like the other former employees, spoke on condition of anonymity because of nondisclosure agreements with Orig3n and fears of retaliation. “Marketing was the priority.”

Orig3n said in a statement that the accounts of its former employees are “grossly inaccurate,” and broadly dismissed the workers as disgruntled. “In some cases, former employees are former employees for a reason,” Chief Executive Officer Robin Smith said in an interview at Orig3n’s glossy offices on the outskirts of Boston’s Seaport District. “We’ve found after employees are gone that they have not done things appropriately.” In its lab, Smith said, the company follows standard federal protocols for analyzing samples and manually reviews results to ensure their accuracy. He said Orig3n now uses lab techniques that don’t require retesting genes already on file, and he believes there were no errors when the company still employed the older testing method under which the former lab tech claims to have logged the 407 errors in a sample of fewer than 2,000 tests. He said that what the lab tech described reflected the use of software meant to prevent employees from manipulating results.

Orig3n isn’t a major player in the $300 million (and growing) market for consumer DNA testing led by 23andMe and Ancestry.com. But it’s raised more than $50 million in five years from big names in the field—including LabCorp, a leading clinical testing company. It also sponsored sports teams such as the San Francisco 49ers. And because it stops short of claiming its tests could diagnose disease risks, Orig3n has been subject to little regulatory oversight and few consequences. Unlike 23andMe, which has received approval from the U.S. Food and Drug Administration to assess customers’ risks of cancer and other diseases, lifestyle DNA tests like Orig3n’s have “the accuracy of a Magic 8-Ball” even at their best, says Laura Hercher, director of student research at the Sarah Lawrence College Human Genetics Program in Bronxville, N.Y.

In the interview, Smith said the tests “gave me a little instruction manual for how to treat my body.” He’d always assumed his shortness of breath limited his ability to exercise, but the tests showed strong muscle efficiency and cardiac output, so he changed what he ate and how he worked out. (A personal trainer helped with that.) He said he can now complete trail runs he couldn’t have imagined before, including a recent 8-kilometer run in California’s Santa Cruz Mountains. LabCorp declined to comment for this story beyond confirming its investment, and the 49ers didn’t respond to requests for comment.

Founded in 2014, Orig3n first billed itself as a regenerative medicine company that planned to build the world’s largest stem cell bank. By 2016 it had shifted its focus to consumer DNA testing. It doesn’t need FDA approval to claim it can assess how well people metabolize fats, what hair products are best for their genes, or whether they’re likelier than the next person to become addicted to drugs or alcohol. The scientific community’s understanding of the genetic underpinnings of complex areas such as nutrition and fitness remains tentative at best, says Robert Green, a Harvard Medical School professor who also oversees a genomics research program. “Because DNA is hot, it is imbued with a kind of aura of scientific verisimilitude that people market these lifestyle products with,” Green says.

Orig3n Chief Scientific Officer Marcie Glicksman says the genes the company includes in tests are all rooted in peer-reviewed science. She says the results can help flag health issues that may need attention.

Smith, who has a background in scientific data management, has poured money into high-profile sponsorships and marketing events, setting up a demo test center where the 49ers play and becoming a headlining sponsor of San Francisco’s annual Bay to Breakers footrace. (Runners, like the 49ers fans, got free test kits.) Orig3n tests are sold widely at Target and CVS stores, and the company has attracted mainstream investors including Hong Kong’s Haitong International Securities Co., an arm of a major Chinese brokerage. “We knew them to be passionate, innovative, and, most importantly, of high character,” John Kane, CEO of Bay to Breakers organizer Capstone Event Group, says of Orig3n. Haitong didn’t respond to requests for comment.

A 2018 investigation into home DNA tests by NBC Chicago found that Orig3n mistakenly processed a dog’s DNA, failing to recognize it wasn’t human while other test-kit makers reported that the samples were unreadable. Former employees say the Orig3n lab was sloppy: Multiple samples might be labeled with the same bar code; DNA and blood samples for the stem cell bank could get mixed up or misplaced; the lab didn’t employ controls to ensure accuracy; handling methods could invite contamination; and when a result wasn’t clear, the former employees sometimes made one up. They also say Orig3n ran tests without proper authorization in its lab at the 49ers’ stadium, and that managers regularly compelled them to write positive reviews of Orig3n’s tests on Amazon.com and Google to offset waves of negative feedback.

A federal survey of the lab found that after the dog DNA report, the company changed its testing and quality-control methods. Smith says lab protocols have improved since the company’s acquisition of a federally certified lab in 2017 and that it’s always abided by scientific standards such as using controls. Lab Director Gordon Siek says he’s “more vigilant” than some of his predecessors. When a Bloomberg Businessweek reporter mailed in two different swabs of her saliva for Orig3n’s superhero test, one after the company acquired a federally certified lab, it returned vastly different conclusions. Only 1 of 6 genes analyzed by both tests yielded the same genetic results.

Smith said the company has issued corrections to inaccurate tests, and that he realizes Orig3n’s advice on how customers should use its test results needs to get better. “Sometimes we look at the accuracy of things and go, ‘Man, that’s not working,’ ” he said. “Our approach and our philosophy is to constantly improve the products.” He says the company never processed tests in its lab at the 49ers’ stadium. Orig3n says it sometimes asked employees to solicit reviews from friends and family and that it also welcomed them from employees themselves.

If not outright regulating the genetic tests of companies such as Orig3n, the FDA might allow companies to voluntarily submit their tests to scrutiny for some sort of seal of approval, says Hercher, the Sarah Lawrence professor. “People want minimal regulation in direct-to-consumer genetic testing to open up availability to the broadest number of people,” she says. “It’s very much ‘buyer beware.’ It just seems like that is not sufficient at this point.”