In January 1989, nine months before Jacob Wetterling was abducted, a 12-year-old boy named Jared Scheierl was walking home from a cafe in the central Minnesota town of Cold Spring, when a man in a car pulled alongside him and asked for directions. The man forced Scheierl into his car, drove him out of town and sexually assaulted him. He wiped down Scheierl's snow suit with a cloth or mitten before telling him to put it back on. Then, after returning him to Cold Spring, he forced Scheierl to roll in the snow, presumably to remove any remaining physical evidence.
What the kidnapper didn't consider in 1989 was that someday, advances in DNA testing would make it possible to analyze even miniscule bits of blood, hair, semen and saliva years after they were deposited and match them to their human sources. DNA testing was in its infancy back then and not widely used among police departments. The Stearns County Sheriff's Office, which investigated the case, put Scheierl's clothing — including the snow suit, a T-shirt, and a sweatshirt — in storage.
It was much later, in 2012, long after DNA had become an evidentiary force in solving crimes, that the Minnesota Bureau of Criminal Apprehension (BCA) took samples from the clothing and analyzed them for DNA. By then, the state's premier lab was using a common and well-understood form of testing called STR, or "short tandem repeat," which involves identifying the repeating sequences on a DNA strand that are unique to each person.
The BCA found in 2015 that a sample from Scheierl's clothing matched DNA from a man named Danny Heinrich. That evidence bolstered a request by investigators to search Heinrich's house, where they found child pornography. That, in turn, led to pornography charges that became the leverage to get Heinrich to admit in court Sept. 6, 2016, that he had killed Jacob and assaulted Scheierl.
Investigators had long suspected that whoever assaulted Scheierl also abducted Jacob nine months later. According to the application for the search warrant in July 2015, the BCA stated with confidence that the DNA from Scheierl's clothing matched that from a hair sample Heinrich had supplied in 1990.
Such a match, the application declared, quoting the BCA's findings, "would not be expected to occur more than once among unrelated individuals in the world population."
Yet deriving that finding was likely more complicated than it appeared, because Scheierl's sweatshirt contained DNA from "two or more" people, making the sample what is known as a "DNA mixture." While the testing achieved its goal in the Scheierl case, experts in the criminal justice system nationwide have raised concerns about how mixed DNA is interpreted as the technology used to analyze complex and infinitesimally small samples advances.
Modern DNA testing is more robust — scientists can analyze samples that used to make them throw up their hands. But it's also, in some ways, becoming less reliable, said Jennifer Friedman, forensic science coordinator for the Los Angeles County Public Defender's Office and a noted expert. "We are pushing the boundaries of reliability way beyond what the manufacturers of the kits anticipated or ever thought we would be doing," she said. "Consequently, when we push the limits of the kits, we are getting very questionable results."
In general, here is how the analysis works: DNA is isolated from the cells in a human sample. Then, especially if the amount of DNA is small, it goes through a process of "molecular Xeroxing," whereby particular segments are copied over and over to create the effect of a larger sample. After that, it's put into a machine often called a "genetic analyzer," which feeds results to a computer, which produces a graph. The graph is full of peaks of varying heights, representing alleles, which are defined as any of several forms of a gene.
The alleles, clustered at multiple locations,show the particular DNA makeup of the sample. Most people have two unique alleles at each location on their DNA strand, one from each parent. So two samples from the same person should produce graphs with similar peaks. In other words, they should match.
DNA mixtures, where traces of more than one person are present, are harder to analyze. They require lab workers to pick and choose among clusters of alleles in order to determine how many contributors there are and which markers belong to whom. There is a level of subjectivity in these interpretations. In some cases, there may be an obviously-predominant contributor. But in others, especially with small samples, alleles can drop out or sometimes mysteriously drop in, camouflaging a pattern or creating the false impression of one.
In mixed DNA, analysts are looking for matches in the heights of the alleles, said John Butler, a fellow with the National Institute of Standards and Technology and vice chair of the National Commission on Forensic Science. "But that's where the challenge becomes, trying to pull out or ascertain what the original combination of alleles are from an individual."
The reputation of DNA evidence is such that the public — having watched it free hundreds of wrongly-convicted people — tends to think it's foolproof. But it's not. A sample can be damaged or degraded, even rendered unusable, in any number of ways. Maybe "it's been out in the sun or underwater or exposed to high temperatures that cause the molecule to break into smaller pieces," Butler said. "You may have materials that are present, chemicals that are present in the sample, that cause it to generate what is called a partial profile. In other words, you can't get all the information." Or, you may have a mixed sample or a very small amount of DNA to analyze. "So all those add complexity to the test," he says. "It's not as simple as comparing one to one, two samples to each other."
It can also be tough to analyze DNA collected decades ago, as it was in the Scheierl case. The preservation of evidence, "is one of the biggest challenges that exist with cold cases," Butler said. "Sometimes, even though you may have a general chain of custody — like you know where the sample was — you don't know how it was actually handled. Twenty years ago, someone may have sneezed on it," introducing their own DNA and creating a mixture where originally there was none. "Now, police officers and technicians that are working with a sample, crime scene examiners, they may be actually doing it with gloves and masks and everything else. But something from 20 years ago, they may not have been as careful because they didn't realize the sensitivity of what can be done with DNA testing."
Established in 1990, a year after the Scheierl assault and the Wetterling abduction, the BCA's DNA laboratory was one of the first in the country. Soon after, it was out front in making a "cold hit" match between DNA found at a crime scene and that contained in a sex offender database. DNA evidence has risen in eminence in Minnesota to the point where the Legislature, in 2000, removed the statute of limitations for serious sexual assaults where testable DNA is present.
But, as analysts have become able to parse ever more-difficult bits of DNA, the possibility of drawing wrong conclusions has increased. "It becomes a two-edged sword," Butler said. "You have the capability of getting more information but you also have the capability of getting that information from other sources." In other words, an analyst might detect DNA from a person who touched a doorknob days before, a person whose DNA essentially hitchhiked into the crime scene on the hand of the perpetrator.
In the past, before current technology, analysts simply didn't interpret mixed samples because they were too complicated. But now, the pressure is on to deliver results. "Law enforcement has really been pushing and district attorneys' offices and government prosecuting agencies have been pushing for DNA from everything," said Friedman. "And this is when the problems really start because laboratories, within their own labs, start coming up with their own rules about how they are going to interpret these mixed samples."
Last year, Washington, D.C.,'s new, $220 million crime lab was shut down after an audit condemned the way it was interpreting mixed DNA samples. The director resigned and operations were suspended for nearly 10 months. There are guidelines for analysis, Friedman said, but no hard standards.
The D.C. lab's crawl back to legitimacy includes the use of new software called STRmix, which is designed to help interpret complex DNA mixtures. Programs like STRmix, and its competitor TrueAllele, represent the next frontier both in analysis and controversy.
Because manufacturers claim their software is proprietary, it's often not possible to replicate the exact process used in an analysis, a hallmark of good science. "It's relatively new," Friedman said. "And frankly, nobody really knows how reliable these computer software programs are. The only publications showing that they are really great are done by the people who developed the programs.
"The only real question at this point is which of the competing software programs the laboratories are going to embrace," she said, and whether the courts will deem their results admissible. "There is going to be a lot of litigation around them," she said.