When traditional DNA testing hits a wall, police are increasingly turning to complex computer programs to sort out whose DNA is in the mix. The U.S. Court of Appeals for the Third Circuit has just decided that one of those tools is reliable enough for admissibility in a criminal case, so long as there’s solid scientific support behind it.
When Old‑School DNA Isn’t Enough
Pennsylvania State Police executed a search warrant at a home where Hunter Ryan Anderson was present while on state parole. In a bedroom, police officers found a bag that held a handgun, two loaded magazines, and Anderson’s identification card, tying him closely to the firearm. Investigators swabbed the gun for DNA and, under a separate warrant, collected a DNA sample directly from Anderson to see if he had handled the weapon.
The state crime lab analyzed the swab and found that the DNA on the gun came from more than one person, which made it hard to interpret using traditional, manual methods. Lab analysts were unwilling to say, “within a degree of scientific certainty,” whether Anderson’s DNA was in that mix.
Anderson was later charged with being a felon in possession of a firearm under federal criminal statute 18 U.S.C. § 922(g)(1). Because that charge ultimately turns on whether the defendant actually possessed the gun, the government wanted stronger evidence tying him to the weapon. So, law enforcement sent the DNA data to a private company that uses specialized software to crunch complicated DNA mixtures and estimate how likely it is that a particular person contributed to them.
The company’s software compared the DNA mixture from the gun to Anderson’s sample and calculated a “likelihood ratio,” a statistic that measures how much more likely the evidence is if a particular person contributed to it rather than a random person. In Anderson’s case, TrueAllele reported that the mixture was 11.5 trillion times more likely if he was a contributor than if some other Caucasian person was. The government treated this as strong forensic support for its theory that he had handled the gun.
Putting the Software Through Daubert
Anderson’s lawyers, of course, moved to keep the private lab’s DNA evidence out of his case. They asked the district court to throw it out under the basic rules that govern expert witnesses in federal cases. One of those rules, often called Rule 702, says judges have to act as gatekeepers. Before a jury ever sees expert testimony, the judge must decide whether the expert is using a reliable method and applying it properly.
This is often called a “Daubert” inquiry because of a 1993 Supreme Court case, Daubert v. Merrell Dow Pharmaceuticals, that spelled out how courts should do it. In Daubert, the Court told judges to consider things like whether a method can be tested, whether it has been tested, whether it has been reviewed by other scientists, what the error rate is, and whether people in the relevant field generally accept it.
Here, defense counsel argued that TrueAllele’s way of analyzing DNA mixtures wasn’t trustworthy enough to be presented in a jury trial, especially since its code is proprietary and the defense couldn’t see how the program actually generated that 11.5‑trillion‑to‑1 number. The judge responded by holding a two‑day hearing focused entirely on the software, with experts on both sides. After that, the trial court decided the government had shown enough scientific backing for TrueAllele and let the DNA evidence in, leaving the finer‑grained attacks for cross-examination rather than excluding the software altogether.
From there, the case moved up to the Third Circuit, which had to decide not just whether TrueAllele could come in, but what that meant for Anderson’s conviction and sentence.
Appeals Court Embraces the Algorithm
On appeal, the circuit court treated the fight over TrueAllele as a familiar expert‑evidence problem: did the trial judge misuse his “gatekeeping” role when he let this kind of DNA testimony go to a jury, or was the software reliable enough to clear Rule 702’s bar? The panel said the judge got it right.
It started by looking inward, to a 2004 decision called United States v. Trala, where the court had already approved a DNA method that relied on software and statistics, even if it wasn’t as sophisticated as probabilistic genotyping. That older case, the panel said, showed the Third Circuit has long been willing to admit scientifically grounded, computer‑assisted DNA analysis when it meets Rule 702. Probabilistic genotyping, in their view, is simply the next step along that same scientific path.
The judges then looked sideways to another appellate court. In United States v. Gissantaner, the Sixth Circuit upheld a different probabilistic genotyping program, STRmix, under the same Daubert framework. That opinion doesn’t bind the Third Circuit, but it does serve as persuasive authority and as a sign that other courts see these tools as part of mainstream forensic science, not experimental outliers.
With that backdrop, the panel walked through the usual Daubert guideposts. The court emphasized that Rule 702 doesn’t demand perfect science; it requires methods that can be tested, have been tested, have known error rates, and are generally accepted in the relevant field. TrueAllele, the court said, clears that bar. The record showed dozens of validation studies, a tiny false‑positive rate (especially at sky‑high likelihood ratios), and broad use in forensic labs.
The judges were not persuaded by the defense’s complaints about hidden source code and the judgment calls made by human analysts. They saw those as reasons for jurors to be cautious about how much weight to give the DNA evidence, not reasons to bar it altogether. In the panel’s view, Anderson got what Daubert promises: a full‑blown hearing where he could test and challenge the method in front of a judge.
The Result
We’ll briefly note that Anderson also argued that 18 U.S.C. § 922(g)(1) violates the Second Amendment, but the Third Circuit rejected those claims and held that disarming felons remains consistent with the historical tradition the Supreme Court requires. The panel also left Anderson’s 78-month sentence in place, affirming the judgment across the board.
What does the court’s decision mean for future DNA cases? At least in the Third Circuit, probabilistic genotyping like TrueAllele is now treated as a form of forensic evidence that can be admitted so long as the record shows solid scientific support. Defense efforts to force disclosure of proprietary source code through evidentiary motions will likely face stiff headwinds. Yet another chunk of the country has just signaled that robust validation and low error rates can be enough, even when the software’s inner workings remain under the hood.
Related Resources:
- Post-Conviction DNA Analysis (FindLaw’s Learn About the Law)
- Crime Scene Forensics: Familial DNA Searches (FindLaw’s Learn About the Law)
- U.S. National DNA Database System (FindLaw’s Learn About the Law)
