Israeli geneticists conducted a curious "investigative experiment" that showed that the identity of an arbitrary US citizen can be established from one DNA sample in 60% of cases, using private genomic databases. Their findings were presented in the journal Science.
“We can say that in the near future, genomic databases will work as a 'GPS system' to search for anonymous owners of this or that DNA. Family trees will play the role of coordinates in it, allowing you to find certain people through their relatives, even if they themselves have not passed such tests,”says Yaniv Erlich from Columbia University in New York (USA) …
Genomic portrait
The development of genomic technologies and the reduction in the cost of DNA decoding procedures have made genetic examination one of the main tools of criminologists, historians and many other specialists not directly related to biology. Today, genomes are used to search for criminals, missing people and reveal the secrets of the origin of peoples.
What's more, last year Craig Venter, a renowned bioentrepreneur and geneticist, said his team was able to find the DNA regions that govern facial shape and other anatomical features. Their analysis, according to the geneticist, makes it possible to draw up the correct composite of a person in 75% of cases. Venter's ideas drew a storm of criticism from other biologists, including Ehrlich.
As Ehrlich noted then, the whole point of this "discovery" was that a person's age, as well as his gender and ethnicity, can be calculated from his DNA, and this data can be used to narrow the circle of potential "suspects." This works for small groups of people, but will not work at the country and large city level.
Such reasoning and arguments with Venter prompted Ehrlich to create a technique that would make it possible to really identify the identity of a random person on the street or help the police search for criminals throughout the country using only single samples of their DNA.
Promotional video:
Today, as Ehrlich notes, companies such as 23andMe, Family Tree, Ancestry and their other competitors are developing especially rapidly, calculating family ties between their customers and determining their predisposition to various diseases based on their DNA samples.
The services of such startups are now used by millions of people in the United States and other developed countries of the world, thanks to which they have accumulated some of the largest genetic databases in the world. Their data is now being used by scientists to search for genes associated with rare hereditary diseases, as well as for many other purposes.
New opportunities and threats
Ehrlich and his colleagues used one of these databases, compiled by MyHeritage, to test whether they could be used for "forensic" purposes - to search for unknown individuals whose appearance nothing is known.
In total, over 1.2 million people have used the services of this startup, many of whom are relatives of each other. Using random DNA samples from people not tested within the walls of this company, Ehrlich and his colleagues checked whether their relatives could be found and "calculated" by matching genome segments.
As it turned out, this can be done for about 60% of Americans of European descent, and in many cases scientists have been able to identify not only second cousins and other distant relatives, but also direct relatives.
Moreover, the calculations of Ehrlich and his team show that a database, which includes only 2% of the inhabitants of a particular country or city, is sufficient to determine the identity of virtually all of its inhabitants, using the same information about their gender, age and eye color and other features that Venter and his associates relied on.
For example, for all 30 “anonymous” whose DNA was analyzed by scientists, the initial list of “candidates” for their role included approximately 800-900 people. When geneticists took into account information on their age, sex and the point where the sample was obtained, they were able to reduce their number to 1-2 individuals.
Such a successful implementation of this "investigative experiment", as Ehrlich notes, speaks of two things. Firstly, law enforcement officers can now safely use genomic databases to find criminals and relatives of their victims. Second, genomic startups should pay much more attention to protecting the personal data of their customers than they do today, and use cryptography to protect them from deanonymization.
