Twists and Turns of Justice

It’s a scenario worthy of a television crime drama: a detective picks up a glass used by a murder suspect and takes it to the lab for DNA analysis. The lab matches the DNA to that found at the crime scene. Voilà, case solved.

Not surprisingly, real life criminal investigations aren’t that simple. As Sheldon Krimsky writes in his new book, Genetic Justice (Columbia University Press), DNA evidence isn’t always reliable. And there’s another red flag: government agencies have created large DNA databases that pose a threat to civil liberties, says Krimsky, a professor of urban and environmental policy and planning in the School of Arts and Sciences.

What interests Krimsky most, he says, is whether “an increasing reliance on DNA in our criminal justice system has furthered our pursuit of justice.”

And justice, he notes, is not the same as solving crimes. We must, he says, balance the harm to individual rights in current practices like collecting DNA without a warrant and retaining the DNA of the innocent in a national database with DNA’s usefulness in identifying criminals and exonerating the wrongly convicted. In the end, Krimsky advocates stronger legislation to protect the rights of the innocent when it comes to the collection of DNA evidence.

He points to “the fallibility of the basic science of the test, the potential for abuse of the information, misjudgments both legal and scientific and the trampling of our rights of privacy and our right to fair treatment in the justice system.”

When DNA testing was first developed, FBI forensic scientists chose to focus on 13 DNA regions, or loci, on the chromosome because they are where the most variability occurs between individuals. These 13 loci are used across all DNA testing to assure consistency in test samples.

Using the output from a DNA analyzer, scientists read and interpret 26 data points at 13 sites on different chromosomes to create a DNA profile of an individual. Crime scene profiles can be compared with those of a suspect to see if they match. The analyzer counts the number of repeated chemical sequences (called short tandem repeats or STRs) at the chromosome sites and produces a graph that reveals the number of STRs at each site. Can two people have the same STR profile?

To achieve 100 percent probability, you would have to compare 3 billion base pairs of the crime scene sample and a suspect’s DNA, not just 26, write Krimsky and the book’s co-author, Tania Simoncelli, a former science advisor to the American Civil Liberties Union.

Instead, what is most often achieved is a random probability match. For example, a crime-scene sample could show 10 STRs and a suspect’s sample six STRs. The question becomes how many others in the population might have this reading—one in 10, one in 10,000? The accuracy of a random probability match, Krimsky says, becomes dependent on a number of factors including, in some cases, a laboratory technicians’ subjective analysis. The same sample might be read quite differently by two separate labs, especially if it involves mixtures or degraded samples. Accuracy of random matches therefore varies from lab to lab and case by case.

“It really breaks down to just 26 numbers,” Krimsky says. “You and me and everyone else, it is 26 numbers that define us when it comes to forensic DNA.”

Chain of Evidence

The mishandling of DNA evidence is another concern, and Krimsky cites numerous examples of errors that have occurred.

In 2002, a 26-year-old man was mistakenly charged with multiple felonies in Las Vegas because the crime lab inadvertently switched two DNA samples. He faced life in prison and was incarcerated for a year before the error was discovered.

In 1993, Timothy Durham was convicted of raping an 11-year-old girl and sentenced to 3,000 years in prison, despite having 11 alibi witnesses who placed him in another state at the time of the crime. In fact, a lab had failed to properly separate the male and female DNA from a semen sample, and the combination produced a genetic profile that could have matched Durham’s. He spent four years in prison before being released.

Krimsky also cites research work at one DNA lab, in Bakersfield, Calif., which he says is indicative of a larger problem. At the lab, a mother’s sample was contaminated by her child’s DNA; samples were accidentally switched or mislabeled; DNA from two samples was combined into one tube; and a suspect tested twice did not match himself.

“Contamination of samples is a huge problem,” says Krimsky, “and labs make mistakes, and people have been sent to prison based on those mistakes. Whenever you have humans interpreting information, there are bound to be errors.”

DNA in a Database

In the last 15 years, the collection and use of DNA by law enforcement agencies have expanded tremendously, Krimsky says. It has become as routine a part of investigations as dusting for fingerprints. Samples can be collected from an individual only with his or her consent or, in more rare cases, a court order. But, Krimsky claims, consent is often given when authorities imply that without cooperation, a person would be considered a suspect.

DNA evidence can also be gathered from a suspect’s home, trash or other belongings without permission or prior knowledge of the suspect. The decision on how or when to collect DNA is left up to investigators.

By March 2010, more than eight million Americans had their DNA collected and retained in a national DNA database, called the Combined DNA Index System (CODIS), which is maintained by the FBI, Krimsky writes. In some states, DNA can be included in the database because it was taken during an arrest, even if the charges were dropped or the person found innocent.

It can also be taken during so-called DNA dragnets—community-wide  samplings of individuals who might be associated with a crime, even though there is no evidence of suspicion of individuals—or familial searches, when a victim’s DNA is a close but not exact match of  a forensic DNA profile  in CODIS. Even if that individual is clearly not the perpetrator, police begin looking at relatives as possible suspects.

Krimsky cites one example of a familial search in which 70 family members consented, when asked by police, to give DNA to the databank when a relative’s DNA partially matched that found on a victim. He puts it this way: “The fact that a witness saw a man with a black hat commit a crime does not enable police to invade the privacy of all men with black hats.”

Familial searching “creates a back-door way for law enforcement to investigate people and their families without their knowledge or consent, let alone a search warrant, because someone in the family has their  DNA in CODIS,” he says.

The same questions arise when police use DNA dragnets to solve crimes, Krimsky says. As he writes in Genetic Justice, when the freelance journalist Christa Worthington was murdered in 2002 on Cape Cod, police used a DNA dragnet to identify a suspect. Three years after the murder, with no leads on a DNA match to that found at the crime scene and the case going cold, they collected DNA from 800 men living in the area, trying to match the sample found at the murder scene. In the end, it was other evidence that led investigators to a trash collector who was not part of the dragnet; he was convicted of the crime.

And the DNA from those 800 men? It all remains in CODIS, even though none was charged with any crime, past or present.

According to Krimsky, being included in CODIS means a person could be screened as a suspect in any future investigation. Worse still, he says, is that no one has clear legal rights to remove his or her biological sample once a case is closed. Under current law, it is incumbent on an individual to petition the court to have his or her DNA removed from CODIS—often a costly and lengthy process, he says. He advocates changing the law so it is incumbent upon law enforcement to remove the DNA of those acquitted or never charged with a crime.

To be fair, it’s not all bad news. Take the case of the Grim Sleeper, a serial killer who murdered at least 10 African-American women in south Los Angeles between 1985 and 2007. He was identified, brought to trial and convicted when a partial match was made between DNA found on all 10 victims and that of his son, who had been convicted of an unrelated felony, resulting in his DNA being added to CODIS.

“Much remains to be done to establish this powerful technological tool in a manner that conforms to our sense of values and justice,” says Krimsky, “and balances the values of solving felony crimes, freeing innocent and wrongly convicted individuals and preserving our rights of privacy.”