MOUNTAIN VIEW — In less than a decade, biotech company 23andMe has turned a refrigerator full of spit into one of the largest databases of personal genetics information in the world.
The brainchild of Anne Wojcicki, the wife of Google co-founder Sergey Brin, 23andMe began in 2006 as a startup mailing DNA testing kits to customers’ front doors and asking them to mail back a vial of saliva. Eight years later, the company is the gatekeeper of a database of hundreds of thousands of people’s DNA — a self-described Google for genetics information.
“It’s actually bigger than anything else I can think of, way bigger,” said Lisa Brooks, program director of the National Human Genome Research Institute, part of the National Institutes of Health.
23andMe has begun selling that genetics data to researchers and pharmaceutical companies to conduct large-scale medical studies, making it an emerging leader in a largely underexplored, and at times hotly debated, area of scientific research. In the last couple of months, 23andMe has announced a joint effort with Pfizer to research inflammatory bowel disease, released findings from a joint study of more than 100,000 people that made new discoveries on Parkinson’s disease, and received a $1.4 million grant from the NIH.
But as the guardian of a very lucrative set of data — the accuracy of which has come under question — critics say the Mountain View company also may pose a threat to consumers’ privacy.
Most medical studies take months or years to solicit enough volunteers. But 23andMe puts the genetic information of 700,000 people at researchers’ fingertips, allowing medical studies to be fast-tracked and new treatments to make their way into hospitals sooner, experts say, giving patients with chronic diseases a better quality of life.
“Instead of actually having to do clinical trials the old-fashioned way, we can enable researchers to get their answers instantaneously,” Wojcicki said in an interview with this newspaper. “And they pay us for that.”
But some experts worry 23andMe users have no idea where their own genetic information will end up. Because the company is relying on data sales to become profitable — selling $99 home genetic testing kits doesn’t pull in the big-dollar revenue — 23andMe may disseminate consumers’ genetic information not only to government agencies and research institutions, say legal and bioethics experts, but also to big pharmaceutical companies, marketers and advertisers.
“There are a lot of people who would want to use that data. There’s a lot of money potentially locked up in that data,” said Charles Seife, a professor at New York University and longtime science writer.
Indeed, in a 2013 interview with The New York Times, Wojcicki said, “I remember in the early days of Google, Larry (Page) would say, ‘I just want the world’s data on my laptop.’ I feel the same way about health care. I want the world’s data accessible.”
Some research experts also question the accuracy of 23andMe’s data, which it collects through surveys it sends to customers, asking about their family and medical histories.
The validity of these studies “is going to depend on the accuracy of the medical information in 23andMe’s database,” said Dr. Paul Appelbaum, director of the Division of Law, Ethics and Psychiatry at Columbia University. “The information 23andMe has is exclusively self-reported, so how accurate it is will become a critical point.”
Regardless, the NIH and the companies buying the data don’t see a problem: “It doesn’t need to be perfect to be helpful,” said Brooks of the NIH. “23andMe has managed at no cost to the government to amass a very large set of data on people.”
23andMe, which has raised more than $126 million and has the backing of Google’s venture arm and investments from Brin, as well as Russian billionaire and Facebook and Twitter investor Yuri Milner, has proved a much more powerful research engine than almost any public effort to understand human genetics, which tend to suffer from underfunding. By comparison, a national genetics study in the United Kingdom is collecting DNA from 500,000 people, and the NIH genome project has 1,000 genome samples.
The NIH grant will allow 23andMe to further expand and refine its database so that it will eventually become a portal of genetics information researchers can access with a keystroke.
The boost from the NIH followed a protracted battle with the Food and Drug Administration, which in November compelled 23andMe to stop offering some of its genetic testing services after regulators questioned the accuracy of the results and warned of the danger of consumers receiving life-changing health information that doesn’t come from a doctor. By that time, 23andMe had the DNA of 550,000 customers.
Customers must sign a consent form for 23andMe to use their genetics information in medical research — and about 85 to 90 percent of customers do, Wojcicki said, explaining that most customers suffer from or have an interest in a genetic disease and want to contribute to scientific research.
But Appelbaum questions how many customers fully understand what they are agreeing to: “That consent rate is quite high. Most studies conducted in academic settings have much lower consent rates,” he said. “That raises the question of how well 23andMe describes what it is that they are asking permission for.”
The company sells genetics data for research only in aggregate “to minimize the possibility of exposing individual-level information,” according to the privacy policy. But customers can also agree to share their personal information, and when given permission, 23andMe will give out details such as their name, email, height, eye color and birth date. The company could not say how many of its customers agree to that level of disclosure.
The privacy policy doesn’t limit with whom 23andMe can share data and stipulates the company can “enter into commercial arrangements” with other companies for the purpose of selling or offering products and services.
Spokeswoman Catherine Afarian said 23andMe asks customers “in every single specific instance in which we want to use personal information” and “there is no scenario in which someone could give us a blanket permission. We believe you own your information. It’s your data, you should have access to it and you should have control over it.”
But in a 2012 study from the Massachusetts Institute of Technology, researchers proved they could determine the identities of some customers of two online genetics testing companies simply by studying their DNA samples, tracing their ancestry to figure out their last names and doing an Internet search. That led researchers to call for stricter data-sharing policies for companies such as 23andMe.
Wojcicki isn’t surprised or fazed by the criticism: “23andMe is forging new ground,” she said. “I’m not here just to try and make a quick buck. We’re code-breaking here, and it’s a really complex problem.”
Contact Heather Somerville at 510-208-6413. Follow her at Twitter.com/heathersomervil.
