These breakthroughs came as a result of an investigative technique called “genetic genealogy,” a blend of DNA analysis and old-fashioned archival research used to point investigators in the direction of a person of interest in a criminal case.
For the past few decades, forensic DNA matching techniques have been used to make direct matches between a sample taken at a crime scene and a suspect who has been identified in a database. “But if you didn’t get a match, the DNA couldn’t tell you anything else,” said Ellen Greytak, director of bioinformatics at Parabon NanoLabs, a DNA technology company based in Reston, Va. “What we’re doing at Parabon is essentially saying: There’s a lot more information in DNA.” Parabon NanoLabs, with genealogist CeCe Moore as head of its genetic genealogy unit, has been at the forefront of the emerging field.
The linchpin of genetic genealogy is actually a simple Internet search: An investigator uploads anonymous genetic data from a crime scene into an open-access genetic genealogy database, such as GEDmatch.com for example, and runs a comparison.
GEDmatch was created to help people learn more about their family trees. To use the site, people download the results from their consumer genetic test and upload that data voluntarily to the database, deciding how much personal information to attach to that data, if any. GEDmatch’s search tool compares one set of genetic data against more than 1 million others in the database, looking for enough shared DNA with other users so that the similarities couldn’t happen by chance. The more closely two people are related, the more DNA they share. Siblings share half of their genes, while first cousins share an eighth. The third cousin level is a sort of red line, past which it’s difficult to tell how closely two people are related, because the amount of DNA shared is so small.
Once the results are in, getting from that initial match to a person of interest takes a lot of legwork. It requires a genealogist searching through birth records, newspaper archives and social media to build out an extensive family tree based on a common ancestor. Then researchers follow branches of that tree forward in time to the present day, to find the right person who was in the area of the crime and whose age and physical characteristics match the description of the perpetrator.
Parabon says it has helped solve more than 30 cases. “These are cases that it’s entirely possible never would have been solved without the information that we’re giving to them,” Greytak said, “and we’re always giving those detectives something they didn’t have before, which is really what is the challenge in these cold cases.”
The results of genetic genealogy can be alarming. According to a recent study published in Science magazine, 60 percent of searches on voluntary consumer genetic databases for people of European descent could identify a third cousin or closer relative. But that match is just the start of an extensive, painstaking process to find the real identity of anonymous DNA.
“There are many misconceptions about privacy and law enforcement use of GEDmatch” said Curtis Rogers, who co-founded GEDmatch in 2010. “It is not your father’s forensics. Yet people are locked into what they see on TV and, to be frank, with rare exceptions it is not law enforcement that use GEDmatch. They need the services of an experienced genealogist and this is where people such as Parabon come into the picture.”
In a response to customer unease about how their genetic data is being used, consumer genetic testing companies such as Ancestry, MyHeritage and 23andMe have signed on to privacy best practices that include being transparent about how many law enforcement requests they get and how many they agree to. A recent survey from researchers at the Baylor College of Medicine found that a majority of the public supports the use of these databases in police investigations. But that support diminishes significantly when the searches are related to cases with nonviolent offenses. In 2017, Ancestry received 34 law enforcement requests and provided data in 31. All were for credit card and identity theft cases.