Barlow points to the findings of little-known Arizona state crime lab analyst Kathryn Troyer, after tests she ran found a pair of genetic profiles that matched, according to state standards. The convicts that her system matched were very much unrelated, however: one felon was white, and the other was black.

Neither state databases, nor the FBI national DNA database called CODIS, keep complete records of a person’s DNA. Instead, databases store a tiny slice of a person’s genetic map – data on 13 specific locations, or loci, along a sample’s (blood, hair, etc.) chromosomes. In the recent past, DNA samples are considered a match when at least 9 out of the 13 loci are found to be identical, however many states now insist on a complete, 13-for-13 match instead. (A lower standard is still applied in cases of damaged, contaminated, or old crime scene evidence.)

Intrigued about her discovery, Troyer and her colleagues ran additional tests and published their findings in “a simple poster” at a national conference of DNA analysts. A handful of attendees said they saw similar findings in their own labs.

Barlow stumbled across the poster three years later, while conducting research for a client accused of 20-year-old charges of rape and murder. Her client was caught by a nine-loci match of his DNA profile and semen found on the victim’s body.

After speaking with Troyer, Barlow obtained a court order to run a radical new search on Arizona’s DNA database, one that would compare every DNA profile stored against every other DNA profile. The results were astonishing: in a database of 65,000 felons, Arizona’s system found nine-loci matches between 122 pairs of profiles. 20 pairs of profiles shared 10 loci, and one pair match at both 11 and 12 loci.

While many of those matches could be explained away – felons in the 11 and 12 loci matches were later determined to be relatives – a heavy majority of DNA matches had no explanation.

Troyer’s findings have the potential to turn the criminal justice system upside-down, bombarding courts with requests to have the supposedly “infallible” DNA evidence reexamined.

The FBI continues to assert the accuracy of its claims, and calls Troyer’s findings the result of an unorthodox, unauthorized style of search. Normal criminal cases do not seek to compare masses of records against each other, and the odds of a false match in the FBI’s permitted check-one-sample-against-the-database search are still in the billions.

Troyer’s findings and Barlow’s insistence have since resulted in requests to run a similar, “Arizona-style” search against CODIS, as well as numerous other states’ internal DNA databases. In many cases, state DNA databases are linked to CODIS.

CODIS head Thomas Callaghan condemned Troyer’s findings as “misleading” and “meaningless,” and the FBI’s national crime lab has threatened to cut off states’ access to CODIS for if they run similar searches – even if they are court-ordered – against it. A massive search against CODIS could have a variety of ill effects, he argued, including knocking the database offline, overloading the FBI’s systems, or data corruption.

Despite FBI resistance, both Maryland and Illinois judges ordered an “Arizona-style” search on their state DNA databases, together adding another 1,000 nine-loci-or-more matches to the pot. Neither state saw their CODIS access cut, nor did they experience any kind of downtime.

In Maryland’s case, a search conducted in January 2007 against a database of 30,000 profiles yielded three “perfect” matches – 13 of 13 loci found identical. Experts say the matches are probably duplicates, or that they belong to identical twins or brothers. The odds of matching two unrelated people, they say, is 1 in 1 quadrillion. The state of Maryland never investigated further.

“DNA is terrific and nobody doubts it, but because it is so powerful, any chinks in its armor ought to be made as salient and clear as possible so jurors will not be overwhelmed by the seeming certainty of it,” said UC Hastings College of the Law David Faigman, who specializes in scientific evidence.

Speaking in a recent phone interview that the Los Angeles Times described as “cautious” and supervised by her superiors, Troyer said she originally saw her findings as “interesting,” and simply “wanted people to understand [that a false positive] can happen.”

“If you’re going to search at nine loci,” said Phoenix lab director Todd Griffith, “you need to be aware of what it means … it’s not necessarily absolutely the guy.”