As anyone who has ever watched CSI can tell you, DNA is well established as the gold standard when it comes to crime solving. Linking the tiniest shreds of DNA evidence together, forensic investigators are able to provide indisputable proof that ties bad guys to the crime or frees the good guys who were in the wrong place at the wrong time—all within an hour so that the CSI team can get ready for next week’s episode.
If only real life was that easy. While there’s no denying the importance of DNA in successfully identifying criminals or clearing suspects, the actual process of gathering and deciphering DNA evidence is considerably more complicated. Poor-quality samples or partial profiles containing degraded DNA can present huge challenges in reconstructing a crime scene and identifying who was (and wasn’t) there. So too can complex DNA mixtures containing genetic data from more than two individuals, especially if any of the individuals are related.
Beyond the issues inherent in trying to decipher and interpret challenging DNA samples, you have to consider the human factor. Mislabeling DNA samples, misinterpreting test results, or accidentally transferring cellular material or DNA from one sample to another can all lead to false reports of a DNA match.
The bottom line is that while forensic DNA testing is overwhelmingly reliable, it has occasional shortcomings. Like any process that involves human interaction and some degree of subjective decision-making, DNA tests are not now—nor have they ever been—completely error-free.
So where does that leave us when it comes to the relatively recent introduction of sophisticated forensic DNA software that employs sophisticated biological modeling and algorithms to interpret DNA profiles?
Forensic DNA software has greatly improved the ability to interpret low-level, degraded, or mixed DNA samples with a significantly higher degree of speed and accuracy than previously possible. Data from DNA testing can now be uploaded and run through an array of probability models, using more information from a DNA profile. The results can then be compared against a person or persons of interest and a likelihood ratio can be calculated, weighed against coincidence. This is ultimately used to resolve highly complex DNA mixtures previously considered too complex to interpret with a high degree of confidence in its validity.
Not surprisingly, some attorneys are voicing concerns that DNA software is simply too new to be able to rely on the results it produces. Pointing to the Daubert or Frye-Mack Standards, prosecutors and defense attorneys alike have charged that forensic software fails to meets the general acceptance test. They also raise questions about whether it has been validated independently and subject to peer review.
To date, the courts generally have rejected such claims. Two 2017 Michigan cases (Michigan v. Burns and Michigan v. Smith) ruled that one such sophisticated forensic DNA software, STRmix™, satisfies all Daubert considerations—i.e., it has been subjected to rigorous testing, validated subject to peer review, and is generally accepted in the scientific community, as well as federal and state courts.
Similarly, the court ruled in Minnesota v. Edwards that the same software is generally accepted in the relevant scientific community and, therefore, admissible in Minnesota courtrooms, while a U.S. District Court judge in New Mexico determined, in U.S. v. Russell, that STRmix™ “has been tested for the purpose relevant here, that such tests have been peer-reviewed and published in scientific journals, and that its analyses are based on calculations recognized as reliable in the field.”
Those calculations use probability models and Markov Chain Monte Carlo (MCMC) methods employed since World War II, and currently used in everything from computational biology and weather prediction to physics, engineering, and the stock market.
Similarly, as the court decisions note, numerous scientific papers about sophisticated DNA software have been published in peer-reviewed scientific journals. In addition, internal validations have been carried out by all labs in current casework, as required by their accreditation, while both the International Society for Forensic Genetics and the Scientific Working Group on DNA Analysis Methods (SWGDAM) have produced guidelines for validating probabilistic genotyping software.
Another charge leveled by some attorneys revolves around the refusal of some developers to reveal the source code and algorithms used in their software. Attorneys contend that it is impossible to review and challenge the solutions provided by forensic DNA software when there is no explanation or understanding of how the software functions.
To a large extent, the courts have been divided on this issue. In People v. Chubbs, a cold murder case in California dating from the 1970s, the trial court determined that Chubbs was entitled to examine the source code under protective order. That decision was overturned on appeal in 2015, however, with the court ruling that Chubbs’ stated reasons to access the source code, even under protective order, did not outweigh trade secret protections.
Similarly, the court in Washington v. Fair, a Seattle homicide case, ruled in January 2017 that it was “not persuaded that a review of the source code is necessary to determine whether TrueAllele [the forensic DNA software used in the case] is reliable … Testimony in this case from both state and defense experts establishes that scientists can confirm the reliability of TrueAllele without access to the source code.”
In New York, however, a federal judge unsealed the source code for the same software program in U.S. v. Johnson after the investigative journalism organization ProPublica filed a motion arguing that there was a public interest in the code.
While there remains no consensus among the courts, some software developers now disclose their code under non-disclosure agreements. Other DNA software is open source, making its code and algorithms readily available for other users to explore and alter. In either instance, however, reputable developers suggest users take part in an extensive user training to ensure that they not only understand how the software works, but also how to correctly represent the results it generates in court.
Almost certainly, these and other legal questions will continue to be raised about forensic DNA software. It’s the nature of anything new that is introduced into the courts. The bottom line, though, is that sophisticated forensic DNA software represents a tremendous breakthrough which, when properly used, can identify criminal activity and be presented in court proceedings with a high degree of confidence in the validity of its findings.
John Buckleton is a noted forensic scientist who has worked extensively in the DNA field as a member of New Zealand’s Institute of Environmental Science and Research Limited (ESR). Dr. Buckleton is one of the developers of STRmix™, a sophisticated forensic software used to resolve mixed DNA profiles previously considered too complex to interpret.
Suggested citation: John Buckleton, DNA Software in the Courts and the Future, JURIST - Professional Commentary, Mar. 13, 2018, http://jurist.org/forum/2018/03/john-buckleton-dna-software.php.
This article was prepared for publication by Kelly Cullen, a JURIST Section Editor. Please direct any questions or comments to him at firstname.lastname@example.org