Consumer DNA Databases Offer Little Identity Protection, Even For Non-Donors

By Nicholas Gerbis
Published: Thursday, October 11, 2018 - 11:10am
Updated: Thursday, October 11, 2018 - 5:16pm
Audio icon Download mp3 (2.08 MB)

Police are applying the same "DNA triangulation" technique that identified as the alleged Golden State Killer to both cold and active cases.

But research published in the journal Science raises concerns about how easily the data can identify anyone — even people who never submitted their DNA.

Based on genetic information from 1.28 million individuals in consumer genetic databases, the researchers found that they could find third-cousin or closer matches for more than 60 percent of Americans of European descent.

The pool of possible identities could then be reduced using demographic data and other clues.

They say that a database of 3 million — roughly 2 percent of the American population — would enable them to raise that number to 90 percent, and that enough DNA will soon exist in public databases to identify almost anyone.

As such searches look to become standard investigative tools, some have raised concerns about their unclear legal protections.

Lead author Yaniv Erlich is Chief Science Officer at MyHeritage, which provided the study data. He says we need to protect genome data under the federal human research "Common Rule."

"The Common Rule currently says that DNA data is not considered identifiable. And we think that, based on the results of our study, and on the results of previous studies, maybe it's time to revisit this point."

Such data are not covered by patient privacy laws and could be used to identify even anonymous donors.

"What people are expressing when they're disappointed in this is a sense that they didn't know their data could be used in these ways," said Jason Robert, director of the Lincoln Center for Applied Ethics at Arizona State University.

Robert said some user agreements spell out how the company can use customer data. But he added that a lack of transparency might shake people's trust that genomic enterprises — consumer or scientific — will protect their data and their privacy.

"Privacy is one of the values that's going to be at stake here, and it's going to have to be balanced against other kids of social concerns, like for instance good policing and good law enforcement."

Science