Over the past twenty years, DNA analysis has revolutionised forensic science, and has become a dominant tool in law enforcement. DNA evidence is key to the conviction or exoneration of suspects of various types of crime, from theft to rape and murder. However,
the disturbing possibility that DNA evidence can be faked has been overlooked.
Scientists at Nucleix based in Tel Aviv, Israel have demonstrated that it is possible to fabricate DNA evidence, undermining the credibility of what has been considered the gold standard of proof in criminal cases.
The scientists fabricated blood and saliva samples containing DNA from a person other than the donor of the blood and saliva. They also showed that if they had access to a DNA profile in a database, they could construct a sample of DNA to match that profile
without obtaining any tissue from that person.
The scientists fabricated DNA samples two ways. One required a real, if tiny, DNA sample, perhaps from a strand of hair or drinking cup. They amplified the tiny sample into a large quantity of DNA using a standard technique called whole genome amplification.
Of course, a drinking cup or piece of hair might itself be left at a crime scene to frame someone, but blood or saliva may be more believable.
The scientists took blood from a woman and centrifuged it to remove the white cells, which contain DNA. To the remaining red cells they added DNA that had been amplified from a man's hair.
Since red cells do not contain DNA, all of the genetic material in the blood sample was from the man. The authors sent it to a leading American forensics laboratory, which analysed it as if it were a normal sample of a man's blood.
The other technique relied on DNA profiles, stored in law enforcement databases as a series of numbers and letters corresponding to variations at 13 spots in a person's genome.
From a pooled sample of many people's DNA, the scientists cloned tiny DNA snippets representing the common variants at each spot, creating a library of such snippets. To prepare a DNA sample matching any profile, they just mixed the proper snippets together.
They said that a library of 425 different DNA snippets would be enough to cover every conceivable profile.