Saturday, April 28, 2018
Chill Factor of a Hot Case
When the news broke on Thursday that a suspect in one of our state's most puzzling cold cases had been apprehended, we found ourselves reading, in the nation's newspapers, terms quite familiar to genealogists. One would hardly have expected news outlets to bandy about phrases like "genetic triangulation," but with Thursday's BuzzFeedNews article on the arrest of suspected Golden State Killer, such specialized terms became mainstreamed.
Arrested this week for crimes committed back in an era when investigators called in their reports from pay phones and wrote them up on a typewriter, the suspect in question had eluded authorities for decades. Had it not been for technological advances developed in the past decade, the man quite likely could have brought his secret with him to the grave.
The game changer was DNA testing. Apparently, investigators associated with a multi-county task force took DNA samples collected from several linked crime sites and uploaded the raw data to a website commonly used by genetic genealogy enthusiasts. From that point, they pinpointed matching relatives, built out family trees, and examined the resultant family members for possible suspects.
Just like we do, when seeking that elusive "brick wall" ancestor. Only this time, the researchers were seeking a rapist and murderer known for vicious attacks across the state of California in the 1970s and 1980s.
There are some problems with this scenario. Early news reports made mention of genealogical websites—signifying use of more than one—as being used in their investigation. Yet Ancestry.com, 23andMe, MyHeritage and Family Tree DNA were quick to announce that they had not been approached formally by law enforcement agencies. Eventually, the news came out that one of the websites that were used was GEDmatch.com, at which time their management posted a statement to that effect on their site.
Even if that was the case, there are still some unanswered questions about the protocol. Those of us who routinely use the DNA testing sites, as well as GEDmatch.com, are quite familiar with how the process works while, admittedly, a journalist might not be as knowledgeable. But one by one, the questions did start trickling through on reports. As a 23andMe representative mentioned in a report by the San Jose, California, Mercury News, "Detectives could not have simply taken the East Area Rapist DNA profile they had from crime scenes...signed up for a service and entered that profile."
In an excellent analysis of the situation, Leah Larkin observed in her blog post at The DNA Geek that it would have been necessary for investigators to "create a mock genealogy DNA test that could be uploaded." Indeed, since commercial companies like Ancestry.com and 23andMe extract genetic profiles from samples sent directly to them by their customers, rather than permitting transfers of raw data to their databases, it would "not be easy for law enforcement to upload a profile to one of those sites," as The New York Times observed. (The only glitch in that assertion, of course, is 23andMe's ironic DNA Day reversal of this policy to allow limited raw data uploads.)
That, of course, would not preclude use of the websites at Family Tree DNA or MyHeritage, which both allow uploads of raw data. Each of these companies, however, has publicly announced that they were not formally approached by law enforcement for such a use.
Essentially, beyond all of the major DNA testing companies providing cousin matches, that leaves GEDmatch. A free, publicly-accessible database, GEDmatch allows uploading of raw DNA data and the subsequent capability of viewing genetic matches from willing others who are also participating in the service.
There were some observations about this use. Besides the question of just how the law enforcement personnel were able to generate the raw data in a format acceptable by GEDmatch (generally, occasional post-update incompatibility with other companies' chips has caused uploading problems for some customers of the main DNA companies), investigators would have had to create a fake user name and profile to gain entrance at GEDmatch. While this is not uncommon—many users try to provide anonymity for the DNA accounts they administer for relatives—the investigators would also have had to check a box on GEDmatch's website certifying that the DNA data they were uploading was either their own or that they had obtained authorization from the person from whom the DNA was obtained. How could they have obtained "authorization" from a person they didn't even know yet?
However it was done, the basic process apparently used by investigators was to upload DNA data to the GEDmatch service, look for family matches, build out a solid family tree for those matches, examine the likelihood of any family members' connection to the crimes (by such parameters as age and geographic proximity), and then zero in on the likeliest suspect.
Once that suspect was identified by DNA data and family tree, there was one more step: find the person in real life, now. Investigators located that one man, set up surveillance, and watched for an opportunity to snag some DNA from him. Within days, they had located an unnamed item the suspect had used and discarded in public. They checked that item for a viable DNA sample and compared the results with DNA gathered from those old crime scenes.
They had a match, initiating the arrest.
The entire scenario has generated an uproar, not only among genealogists but among ethicists and civil liberties proponents. On Friday, a New York Times reporter asked permission to query members of the "Genetic Genealogy Tips & Techniques" Facebook group about their current concerns regarding participation in the GEDmatch website, which generated a lively discussion and resulted in an article published online that same evening.
While some GEDmatch users were quite pleased to learn that techniques they use every day were applied to solving a notorious cold case, many others voiced concerns about the potential for misuse of what has been clearly designated as a set of tools for genealogists, not criminologists. Perhaps instigated by the general angst of our era, an overall sense of mistrust of intrusion by governmental agencies seeped through many of the comments on that discussion thread at Facebook.
One issue in particular seemed the obvious next question: now that we've seen how law enforcement officials can use such databases, what chill factor will reverberate through the future of genetic genealogy? It's hard to say in the heat of the moment, but over the long term, it will likely make its impact known in the aggregate through the individual choices of those who don't care to have their privately-submitted data co-opted by those who are not part of the intended users of this community.