MIX13 Reveals Potential Errors in DNA Testing
by Jayson Hawkins
A federal study from 2013 showed that manually sorting DNA mixtures is not as foolproof as previously believed. MIX13, which sent the same hypothetical cases to 108 crime labs around the U.S., tested the accuracy of traditional DNA analysis. Each of the five cases grew more complicated until the last, which involved a mixture of four individuals’ DNA collected from a ski mask at a robbery. The labs were presented with the identities of two of the likely suspects, along with a fifth person who was not involved.
Just seven labs managed to fully solve the problem; worse, more than 70 percent implicated the fifth “innocent” suspect in their findings.
John Butler of the National Institute of Standards and Technology said the purpose of MIX13 was to show the limitations of using combined probability of inclusion (“CPI”), not to expose the probability of mistakes. “This was a teaching moment to realize you can falsely include somebody with CPI.”
The few labs that correctly answered MIX13 employed rigorous techniques or advanced technology like TrueAllele, a genotyping software.
Critics charge that errors associated with CPI are more than possibilities—they have already happened. The Virginia Department of Forensic Sciences used TrueAllele to take a second look at 144 cases and found five where suspects should have been excluded. Another case in Georgia has been granted a retrial due to a TrueAllele analysis of DNA mixtures from evidence. The defendant, Johnny Lee Gates, has spent over 40 years in prison for a crime he may not have committed.
A paper in Forensic Science International: Genetics criticized the six-year delay in releasing MIX13’s results and demanded that labs begin using the updated technology.
“The adoption of probabilistic genotyping by many laboratories will certainly prevent some of these errors from occurring in the future, but the same laboratories that produced past errors can also now review old cases with their new software—without additional bench work,” emphasized Greg Hampikian, the paper’s author.
More agencies have employed genotyping software in recent years, but the majority still rely on manual methods.
As a digital subscriber to Criminal Legal News, you can access full text and downloads for this and other premium content.
Already a subscriber? Login