The other benefit he sees is the ability to create and store genetic data in-house. The way companies like Parabon and FamilyTreeDNA work is they do the DNA processing, database-searching, and family-tree building, then return a list of names of possible suspects to their law enforcement clients. That kind of work can get expensive quickly, which is why law enforcement agencies with limited budgets have tended to use such services only for very high-profile unsolved crimes.
Hellwig started Intermountain Labs to work on the other cases—the murders that didn’t make headlines, the unresolved missing persons cases, and the nation’s shameful backlog of more than 100,000 untested rape kits. “We’re a nonprofit, so we don’t care who solves the case, we just want to get it solved,” says Hellwig. By generating a genetic profile in-house, he’s free to share that data with volunteer genealogists at the Utah Cold Case Coalition, or with other groups of benevolent genetic sleuths. “No one has a monopoly on the data this way,” says Hellwig.
The steady stream of headlines hailing genetic genealogy as the cracker of cold cases over the past two years has created a new sense of urgency for law enforcement agencies to get in the game. Though there is no centralized count of cases solved with the method across different databases, Williams told WIRED that GEDmatch has so far been used to identify suspects in more than 200 cases. (There’s no tally available yet of how many of those people have been convicted, but it is not zero.) “There is no better motivator than that public pressure,” says Hellwig.
Verogen’s new genetic genealogy product is intended to provide the final push. But it also has features clearly intended to mollify privacy skeptics. The new kits generate a genetic profile that’s much skimpier than what you’d get out of an Ancestry or 23andMe test—about 15,000 data points versus 600,000. Because there’s so much less information, Verogen had to develop a different type of algorithm for turning up potential family members and estimating their degree of relatedness. Rather than comparing the amount of DNA overlap between two people, it now looks for patterns of similarities at single locations proven to be highly predictive of kinship. That means it’s not compatible with the existing profiles GEDmatch has on file for its users, only about 325,000 of whom have consented to law enforcement searches of the database.
So the company has converted those consented profiles into a format suitable for searching with the new DNA kit, while leaving out the profiles of legacy GEDmatch customers who hadn’t opted in. Their team is in the process of creating a separate portal just for law enforcement agencies to upload their crime scene sample data. The idea is to make it technologically impossible for unauthorized searches of non-consented users. (That’s a thing that has happened in the past, not just at GEDmatch, but at several private consumer databases, as The LA Times revealed last week.) Williams says the infrastructure changes should more completely wall off those who do not wish to involve their DNA with crime-solving intrusions from law enforcement.
Further, the DNA test kits don’t scan any portions of the genome known to be medically important, says Williams. “The way it’s been done before is really an appropriation of technology not intended for forensic use,” he says. That created understandable privacy concerns. In addition to family ties, the kinds of genetic data created from consumer spit kits can contain information about people’s looks, ethnic heritage, and medical risks. Though the US has laws preventing employers from discriminating against individuals based on their genes, police have much freer reign with such information. “So what we’re trying to do is step back and say, ‘How do we build something suitable for law enforcement investigators that minimizes privacy violations as much as possible?’” says Williams. “Because the genie is definitely out of the bottle at this point.”