The Surprisingly Imperfect Science of DNA Testing

That’s the title of this in-depth look at DNA testing through the lens of one cold murder case.

Although DNA evidence is widely regarded as the gold standard of forensic science, we should be careful about checking our brains at the courtroom door just because the government exclaims, DNA!

The science works well when we can test an ample, well-preserved sample of a person’s DNA, and when it’s obvious how the sample got to the alleged crime scene, and when investigators and crime labs don’t make mistakes in collecting and testing it.

But that’s not the norm, of course.

The reality is we’re not always able to collect enough DNA to test well. The sample may be too small, or it may have degraded over time, or it may be mixed up with other people’s DNA. In one case, police had collected a partial sample from blood that was found at the scene of a burglary, and it seemed to match the profile of one man definitively. The only problem was that the man suffered from advanced Parkinson’s disease and could barely walk, which made him an unusual burglary suspect. The police arrested him anyway, though, because the odds were supposedly only one in 37 million that it was someone else. Those odds were wrong, however, and later, more advanced testing revealed that the man shared a partial profile with the true culprit.

Nor is it always clear how our DNA ends up where it does. The reality is our cells often find their way into our environment, and sometimes, our DNA winds up in places we’ve never been. In one case, police had discovered the DNA of a homeless man on the fingernails of a murder victim, so they arrested the man even though he seemed to have an airtight alibi: he was unconscious in a hospital bed at the time of the murder. The man ended up spending five months in jail until the authorities realized why his DNA was on the victim’s fingernails. The same paramedics who’d transported him to the hospital had responded to the murder afterward, and they’d clipped the same oxygen-monitoring probe to both persons’ fingers.

Finally, human error can affect the collection, preservation, and testing of DNA evidence at each link in the chain of custody as well as afterward, when analysts may need to interpret test results. While some crime labs are better than others, interpretation begets subjectivity, and subjectivity breeds error. A recent government survey, for example, asked 108 labs to look at a three-person mixture and determine whether a suspect’s DNA was present in the sample. Fully 70 percent of the analysts determined that it could be in the mix, while 24 percent said the data was inconclusive. Just six percent determined correctly that it was not.