3 May 2018 (Rome, Italy) – There was a fascinating conference today here in Rome on how ever-faster sequencing techniques due to machine learning and extraction protocols have been able to bait specific sections of human DNA out of the vast soup of non-human genetic material in ancient samples. Simultaneously, the field has grabbed the public imagination with findings about the distant past. There was a marvellous presentation of the intricacies and difficulties of the field in the DNA analysis of the remains of England’s King Richard III, discovered under a car park in Leicester in 2012. There are now super-sophisticated statistical and bioinformatics techniques available, and it is completely changing our earlier perceptions on how humanity spread across the globe, and uncovering the tremendous degree to which populations actually globally blended, repeatedly, over generations.
But there was also chat about DNA analysis in modern times. Last week’s arrest of a suspect in the Golden State Killer case in California has highlighted how DNA samples that have been volunteered for one purpose – in this case, genealogy – can be used for other reasons, often without the donor’s explicit consent. Several ethicists have expressed concern about U.S. detectives using a genealogy website in this way. Coming so soon after the reuse of Facebook data in political campaigns in the Cambridge Analytica scandal, it’s another example of how new technology and techniques lead to unexpected conundrums, and how ethical and societal debate must catch up.
Briefly, the case of the Golden State Killer, linked to at least 50 rapes and 12 murders between 1976 and 1986, had gone cold – although investigators believed they had a reliable sequence of the perpetrator’s DNA. Next they needed a match. So, according to reports, they uploaded the data to a popular website that compares people’s genetic information to trace their relatives – in effect, creating a profile for him. They got lucky: a match with family members led them to identify and arrest Joseph James DeAngelo.
Just like the Cambridge Analytica case, this one raises the question of how much control people have over information they give to public or commercial databases. DeAngelo’s relatives submitted their DNA for the specific purpose of genealogy, which by definition requires the information to be shared and compared. Then they saw it used for something else without their consent. In discussions of the case, users of genealogy services are divided between those who say the police were justified, given the seriousness of the crimes, and those who were shocked by the move.
Such users have received other surprises. Thousands of people have discovered through genetic analyses that their parents were not who they thought they were. Others have found and been reunited with siblings they never knew existed. Such discoveries have implications for users’ wider family members, most of whom won’t have put their DNA in such a database.
In the California case, the involvement of the police adds an extra dimension. People who choose to upload their DNA could unknowingly be helping police to trace a relative – now and in the future.
In Europe, investigators have long coveted the genetic information held in others’ databases. After the Swedish politician Anna Lindh was assassinated in 2003, Swedish police asked for access to a suspect’s DNA stored in a biobank, so that they could compare it with DNA found at the crime scene. Their access was granted. But other requests have been turned down by courts. In 2006, the Norwegian Supreme Court said that police investigating a suspected armed robber, who had died six months after the crime, could not access his genetic information held by a hospital. The dead can’t be libelled, but they can have their privacy invaded. And scientists in Belgium wrestled with these issues in 2016, when they confirmed the location of the 1934 death of King Albert I from blood samples collected there. They decided not to publish sequence details because of possible implications (including paternity and health) for his surviving descendants, including members of the current Belgian and British royal families.
To what extent can scientists and companies who collect such information anticipate future uses and make them clear to participants and customers? As we are quickly finding out, the technology to develop information – and access to it – is moving at a lightening pace. There are no easy answers, because many of those uses cannot be anticipated at the time. One presenter at the conference today noted that online firms that collate and compare DNA for consumers are often too vague about how it might be used. Critics say these organizations should take steps to inform people better.
Interesting note: in the Golden State Killer case, users of the site that the police accessed, GEDmatch, were told that “other uses” were possible and that they should remove their data if this was unacceptable – with “other uses” not defined.
If police can use genetic databases to catch killers – even those who are distant relatives of individuals who have submitted their DNA – then perhaps more people will sign up to share their DNA. But they should be told that this is a possibility, and be given the choice to opt out. And this year after Facebook, GDPR and other issues you will be hearing quite a lot about “opt out” applied to a myriad of on-line applications. Meanwhile, geneticists, ethicists and lawyers here freely admitted that the law and regulation will – as always – be playing catch-up.