Study: DNA websites cast broad net for identifying people

October 11, 2018 by Malcolm Ritter
Study: DNA websites cast broad net for identifying people
In this Friday, April 27, 2018 file photo, Joseph James DeAngelo, 72, who authorities suspect is the "Golden State Killer" responsible for at least a dozen murders and 50 rapes in the 1970s and 80s, is accompanied by Sacramento County Public Defender Diane Howard, right, during his arraignment in Sacramento County Superior Court in Sacramento, Calif. Authorities said they used a genetic genealogy website to connect some crime-scene DNA to DeAngelo. (AP Photo/Rich Pedroncelli)

About 60 percent of the U.S. population with European heritage may be identifiable from their DNA by searching consumer websites, even if they've never made their own genetic information available, a study estimates.

And that number will grow as more and more people upload their DNA profiles to websites that use genetic analysis to find relatives, said the authors of the study released Thursday by the journal Science.

The use of such databases for criminal investigations made headlines in April, when authorities announced they'd used a genetic genealogy website to connect some crime-scene DNA to a man they then accused of being the so-called Golden State Killer, a serial rapist and murderer.

In general, such searches begin on a site by finding a relative linked to a DNA sample. Then sleuths can use other information like published family trees, public records and lists of survivors in obituaries, plus whatever they know about the person whose DNA began the process. They can build their own speculative family trees. Eventually, that can point to someone whose DNA is then found to match the original sample.

With DNA databases "you need just a minute fraction of the population to really identify many more people," said Yaniv Erlich of Columbia University, an author of the study.

Each person in a DNA acts "as a beacon that illuminates hundreds of distant relatives," said Erlich, who is also chief scientific officer of the MyHeritage website.

His paper focused on Americans of European descent because such people are over-represented in DNA databases, which makes it easier to find relatives.

The researchers started with the 1.28 million participants on the MyHeritage site at the time they did the work. Most had a northern European genetic background. For each, they looked for relatives more distant than first cousins elsewhere in the database.

About 60 percent of the time, they found someone whose genetic similarity was at least equal to that of a third cousin, similar to the degree of relatedness that led to the Golden State Killer suspect. Third cousins share great-great-grandparents.

With some basic assumptions about what kind of data would be available for a criminal suspect, the researchers calculated they could pare down the possible identity of the initial person to just 16 or 17 people. That's limited enough that police could zero in with further investigation, Erlich said.

Erlich and his co-authors suggested that such searches could cast a broader net in the near future. A database with DNA profiles of just 2 percent of a population is enough to match nearly everybody with somebody who's as closely related as a third cousin, researchers said. From that, they calculated that the genetic profiles of about 3 million Americans of European descent could deliver the equivalent of a third cousin for more than 90 percent of that ethnic grouping.

Websites are getting very close to that, said Erlich, noting that MyHeritage now has more than 1.75 million participants. He said the does not allow forensic searches.

Two DNA experts unconnected to the study said third and fourth cousins can both lead to identifications.

Study: DNA websites cast broad net for identifying people
In this Aug. 1, 2014 file photo, tools used for DNA testing are shown in a DNA lab at the forensic science center of the Oklahoma State Bureau of Investigation in Edmond, Okla. About 60 percent of the U.S. population with European heritage may be identifiable from their DNA by searching consumer websites, even if they've never made their own genetic information available, a study estimates. That number will grow as more and more people upload their DNA profiles to websites that use genetic analysis to find relatives, researchers said. (AP Photo/Sue Ogrocki, File)
"Because the average person has so many of these distant cousins, it becomes reasonably probable that one or more of them is in a publicly searchable database, even if only a small fraction of the U.S. population is included," Graham Coop and Michael Edge of the University of California, Davis, wrote in a statement to The Associated Press.

"The fact that most suspects could be identified in this way is predictable" from mathematical calculations, and the new paper provides a convincing demonstration, they said.

However, the work raises important policy questions, they said. Should anyone other than law enforcement be allowed to conduct such searches? And under what circumstances should they be permitted?

"How should we react to the fact that the decisions of our fourth cousins, whom one may never have met, affect one's privacy?" they asked.

In an interview, Edge noted that when people add their DNA profiles to a publicly searchable genealogy site, "they're not necessarily thinking about the genetic privacy of their ."

Amy McGuire, a professor of biomedical ethics at the Baylor College of Medicine in Houston, said that police searches using DNA and genealogy websites have sometimes pointed to an incorrect person.

"You would hope ... the victim of the false lead can be easily cleared" by providing DNA, she said. "But you still have some invasion into that person's personal life by being investigated."

Some people would say that's worth it to aid the cause of justice, but others "would find that very distressing," she added.

McGuire said there's an active legal debate about whether police should be able to "go on a fishing expedition" using DNA genealogy websites without a warrant.

She recently published a survey that suggests most people support letting police search genetic genealogy databases. But support was much higher for investigations involving violent crimes and crimes against children than for nonviolent crimes.

Explore further: How cops used a public genealogy database in the Golden State Killer case

More information: Y. Erlich el al., "Identity inference of genomic data using long-range familial searches," Science (2018). science.sciencemag.org/lookup/ … 1126/science.aau4832

Related Stories

How cops used a public genealogy database in the Golden State Killer case

May 1, 2018
DNA was credited for cracking the decades old cold case of the "Golden State Killer," a California serial murderer and rapist. But the detectives used a public database of genetic genealogy called GEDmatch, raising privacy ...

Computational model links family members using genealogical and law-enforcement databases

October 11, 2018
The notion of using genetic ancestry databases to solve crimes recently crossed from hypothetical into credible when police used an online genealogical database to track down the alleged Golden State Killer, a serial criminal ...

DNA search for California serial killer led to wrong man

April 28, 2018
Investigators hunting for the so-called Golden State Killer turned to searching genetic websites in 2017 but misidentified an Oregon man as a potential suspect. A year later, after using a similar technique, they are confident ...

How DNA led to the elusive 'Golden State Killer'

April 27, 2018
Detectives in California used DNA left at crime scenes, combined with genetic information from a relative who joined an online genealogy service, to catch an alleged rapist and murderer who eluded authorities for four decades.

Bioethicists suggest ethical considerations for forensic use of genetic data

May 28, 2018
Despite the popularity of online genealogy services, it is unclear whether users understand that their genetic information is available for forensic purposes. Bioethicists from the National Institutes of Health (NIH) suggest ...

Investigators say DNA database can be goldmine for old cases

June 16, 2018
A microscopic thread of DNA evidence in a public genealogy database led California authorities to declare this spring they had caught the Golden State Killer, the rapist and murderer who had eluded authorities for decades.

Recommended for you

Scientists identify critical cancer immunity genes using new genetic barcoding technology

October 20, 2018
Scientists at Mount Sinai have developed a novel technology for simultaneously analyzing the functions of hundreds of genes with resolution reaching the single cell level. The technology relies on a barcoding approach using ...

A single missing gene leads to miscarriage

October 19, 2018
A single gene from the mother plays such a crucial role in the development of the placenta that its dysfunction leads to miscarriages. Researchers from the Medical Faculty of Ruhr-Universität Bochum (RUB) have observed this ...

Making gene therapy delivery safer and more efficient

October 18, 2018
Viral vectors used to deliver gene therapies undergo spontaneous changes during manufacturing which affects their structure and function, found researchers from the Perelman School of Medicine at the University of Pennsylvania ...

Student develops microfluidics device to help scientists identify early genetic markers of cancer

October 16, 2018
As anyone who has played "Where's Waldo" knows, searching for a single item in a landscape filled with a mélange of characters and objects can be a challenge. Chrissy O'Keefe, a Ph.D. student in the Department of Biomedical ...

Researchers use brain cells in a dish to study genetic origins of schizophrenia

October 16, 2018
A study in Biological Psychiatry has established a new analytical method for investigating the complex genetic origins of mental illnesses using brain cells that are grown in a dish from human embryonic stem cells. Researchers ...

Why heart contractions are weaker in those with hypertrophic cardiomyopathy

October 16, 2018
When a young athlete suddenly dies of a heart attack, chances are high that they suffer from familial hypertrophic cardiomyopathy (HCM). Itis the most common genetic heart disease in the US and affects an estimated 1 in 500 ...

0 comments

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.