How Courts Decide When Computer DNA Evidence is Reliable

"If you dislike jargon, buckle up," wrote Third Circuit Judge Emil Bove. "The focus of this appeal is the reliability of probabilistic genotyping software in forensic DNA identification under Daubert." – United States v. Hunter Anderson (2026)

Admissible scientific evidence

When one side in a legal dispute wants to call an expert to testify at trial, the opposing side can challenge the reliability of the expert's scientific evidence. If the opposition wins their challenge, the evidence is blocked, and the jury will not hear it.

The Daubert standard provides a legal framework for deciding whether a forensic science technique is sufficiently reliable evidence. A trial judge serves as the gatekeeper, applying the admissibility standard to an expert's methodology. Unlike earlier standards, Daubert emphasizes scientific testing.

On March 26, 2026, the federal Third Circuit Court issued an appellate decision in US v. Anderson. Their precedential opinion affirming the reliability of Cybergenetics TrueAllele probabilistic genotyping software is a model Daubert ruling. The judges clearly explain the science and law, review the defendant's unpersuasive arguments, apply the Daubert standard, and conclude that "TrueAllele is reliable enough to be admissible at trial."   

Anderson DNA results

On Oct. 31, 2019, while executing a search warrant of Hunter Anderson's home in Centre County, Pennsylvania police found a Smith & Wesson .40 caliber firearm. The crime lab found a DNA mixture on the gun whose complexity gave "no interpretable results." As such, "no comparison could be made to" suspect Anderson.

The county's District Attorney's office sent the lab's DNA data to Cybergenetics. The DA had just one question, "Did the suspect contribute DNA to the evidence?" Cybergenetics ran the four-person mixture DNA data through their TrueAllele computer system. The company's initial free case screening connected the gun to the suspect. The DA requested a full report.

Cybergenetics ran TrueAllele on the gun evidence five times, getting consistent results. They reported that a match between the handgun and Anderson was "11.5 trillion times more probable" than coincidence. They also reported an error rate—"for a match strength of 11.5 trillion, only 1 in 146 trillion people would match as strongly."

The Daubert standard

When can courts rely on scientific findings to make judicial determinations? Why should we trust TrueAllele's reported 11.5 trillion match statistic? By what standard can a judge—who is not a scientist—establish the reliability of a scientific methodology?

The U.S. Supreme Court's 1993 Daubert v. Merrell Dow Pharmaceuticals, Inc. decision addresses these questions. The Court ruled that the Federal Rules of Evidence (FRE) enacted by Congress superseded the 1923 Frye general acceptance standard. FRE Rule 702 governs the testimony of expert witnesses, with reliability a key component. Drawing on Karl Popper's philosophy that "falsifiability" confers scientific status, in Daubert the Court shifted scientific reliability from trusting authority to testing hypotheses.

The Court developed five Daubert prongs – testability, peer review, error rate, standards, and general acceptance. A judge can apply these prongs to a proponent's scientific methodology to assess its reliability as evidence. The Court's new theme was empirical testing—a testable method is more reliable. This modern legal standard accords with science.

Defense Daubert challenge

Anderson's defense team challenged TrueAllele reliability. Their motion conceded key Daubert reliability prongs—the software was tested, peer reviewed and published. Moreover, they acknowledged Cybergenetics transparency in having previously let defense experts review its source code, the trade secret text underlying testable software.

There have been over 50 TrueAllele admissibility hearings. Cybergenetics helps trial lawyers prepare and provides hundreds of supporting exhibits. These documents comprised the core Daubert material presented by the U.S. attorney to the Middle Pennsylvania District Court.

Would other probabilistic genotyping (PG) programs also link the gun DNA mixture to Anderson? Why read source code for 10 years when you can simply test the software on data?

We answered these questions by running five publicly available PG programs on the Anderson gun data. The test results are shown in the table below. All the programs connected the gun to Anderson. The two most powerful PG programs (right-most columns) gave match statistics in the trillions, just as TrueAllele had reported.

District Court Daubert hearing

Chief District Judge Matthew Brann held a Daubert hearing "to address preliminary questions relating to the admissibility of [the TrueAllele] expert evidence."  The Judge heard my testimony in December 2022, and then from two defense experts the following month.

Judge Brann's May 2023 Memorandum Opinion reviewed each of the Daubert prongs. He considered Anderson's arguments, but found that the Daubert reliability factors all favored TrueAllele reliability. The judge ruled that "TrueAllele's DNA identification methodology and the match statistic reported in this case sufficiently reliable to warrant admission."

A juror could now hear the admitted TrueAllele gun DNA evidence. But none did. Instead, Anderson pleaded guilty to unlawful possession of a firearm; he was sentenced to 6.5 years in prison. He appealed his conviction. The Third Circuit Court considered his appeal this year.

Other courts deciding TrueAllele reliability have relied on Judge Brann's reasoned Anderson opinion. Jurisdictions include Hawaii (State v. Xavier Swofford; State v. Eric Thompson), New York (People v. Edward Holley), and the federal Middle District of Louisiana (United States v. Damond Lockett). Importantly, the District Court's Daubert reasoning laid the foundation for the later Circuit Court Anderson decision.

Third Circuit Anderson opinion

The Third Circuit Court of Appeals Opinion in US v. Anderson was "a natural extension of existing precedent" that built on the District Court's "thoughtful analysis." The appellate ruling "substantially agreed" with the Sixth Circuit's seminal reversal in US v. Gissantaner five years before. There, Chief Judge Jeffrey Sutton had clarified why different PG software was reliable under Daubert.

The Third Circuit Opinion considered each Daubert prong in turn. It assessed how well the government established that prong for TrueAllele, and addressed opposing arguments.

The following sections parallel Judge Bove's five-prong outline. They illuminate how the Third Circuit Court "addressed Anderson’s arguments on appeal and underscored some of the District Court’s points regarding reliability under the relevant [Daubert] standard." The bracketing Questions and Answers are taken directly from the Opinion.

Daubert prong 1 – Testability

• Question: Is TrueAllele capable of being tested?

The Daubert issue is whether the expert's methodology can be objectively tested, or merely reaches unchallengeable conclusions. "TrueAllele can be tested," wrote the Third Circuit, "by creating multi-source DNA mixtures in lab settings for the software to analyze." Based on such data, false negative and positive outcomes can be observed. That suffices, said the Court.

But the defense didn't test anything, noted the Court, despite being given every opportunity. They had access to the TrueAllele method, a computer for running the software, relevant case and validation DNA data, and testing results. They could have conducted their own scientific testing to challenge the software's methodology. Or to contest the testing of our 42 validation studies.

Instead, they insisted on having TrueAllele source code. Not just to inspect; their experts had already done that for a week in another case. But to possess sensitive trade secret materials for uses unrelated to standard empirical forensic testing.

Why not first use available source code from the other PG programs we ran that all (wrongly, in their view) connected the gun to Anderson? Or, like scientists, test those programs themselves?

The appellate court would "not authorize a fishing expedition through TrueAllele’s source code under the auspices of Daubert. For purposes of testability, what matters is that the challengers of the methodology could run scientific tests to try to show that TrueAllele does not function in the manner that the government’s expert described."  They could have, but didn't.

• Answer: The government established that TrueAllele’s testability supported a finding of reliability.

Daubert prong 2 – Error rate

• Question: Is TrueAllele susceptible to error-rate calculations?

The Court noted that TrueAllele validation studies had established the software's false positive error rate. That its error rate decreases with increasing match strength. And that an "infinitesimally small" error rate of one in 146 trillion was reported in Anderson.

The Defendant argued that "these error rates are inaccurate because, in his view, they fail to account for potential errors by software operators and problems lurking in TrueAllele’s source code." He cited a general error rate of "six errors per thousand lines of code."

"We are left, then," concluded the Court, "with little more than defense speculation that people sometimes make mistakes, source code sometimes contains errors, software sometimes malfunctions, and TrueAllele may suffer from one or more of these issues." But "these concerns are not substantial enough to undermine reliability under Rule 702."  

• Answer: The government demonstrated that it is possible to calculate error rates relating to TrueAllele’s performance, and that the software’s error rates are low. These considerations also favored admissibility.

Daubert prong 3 – Standards

• Question: Is TrueAllele governed by standards that can be applied to reduce errors?

Cybergenetics complies with standards for empirically validating PG software. These national standards were established by the FBI's Scientific Working Group on DNA Analysis Methods (SWGDAM), ANSI/AAFS, and other organizations. At the hearing, the government provided TrueAllele exhibits that document testing compliance.

The Defendant objected that the state crime laboratory hadn't followed PG validation standards for TrueAllele. And that the Institute of Electrical and Electronics Engineers (IEEE) has additional (voluntary) software engineering standards.

But the Court found these arguments unpersuasive. The state laboratory doesn't use TrueAllele, so the PG standards aren't applicable here. And PG software isn't "subject to additional standards" from other fields, "particularly in light of the extent of the existing standards and the weight of the other factors."  

• Answer: TrueAllele’s probabilistic genotyping methodology is governed by the type of standards that are a hallmark of a reliable methodology grounded in hard science. We also care about standards because, if properly applied, they can reduce error rates. That is true here.

Daubert prong 4 – Peer review

• Question: Is TrueAllele a product of peer-review scrutiny?

"Of the 42 TrueAllele validation studies presented by the government at the Daubert hearing, eight of them were subject to peer review," wrote the Court. "One of the peer-reviewed studies addressed TrueAllele’s efficacy with DNA mixtures involving up to 10 contributors, which are more complex than the mixture swabbed from the gun in this case."

The Defendant argued that I (the expert) had coauthored seven of the eight published studies. But the Court found that "of little consequence under Daubert." What's important "is that the peer reviews were conducted by anonymous independent scientists."  

• Answer: TrueAllele has been subjected to peer review and publication. These processes are not “necessary conditions of reliability.” But going through the process is another “component of good science.” (See also Gissantaner: "Publication in a peer-reviewed journal alone typically satisfies this Daubert inquiry.")

Daubert prong 5 – General acceptance

• Question: Is TrueAllele generally accepted in the relevant scientific field of DNA evidence interpretation?

The Court observed that TrueAllele was deemed admissible in 37 cases in the United States. And that I had "testified that TrueAllele had been used in a variety of settings for more than 25 years, including by approximately 400 organizations involved in law enforcement, criminal defense and exoneration work, paternity testing and kinship analysis, and identification of remains."

The Defendant cited a Ninth Circuit non-precedential discussion about a National Institute of Standards and Technology (NIST) report. A draft NIST report had mentioned a supposed lack of "established and accepted criteria" for low-level complex DNA mixtures. The issue didn't appear in their final report. (And is now a moot point.)

The Defendant also claimed that TrueAllele isn't generally accepted in "the community of computer scientists and software engineers." But the Court said that Daubert construes the relevant scientific community more narrowly, here as "forensic geneticists."

• Answer: Probabilistic genotyping is generally accepted, for purposes of Daubert, in the field of forensic DNA identification. This is another "important factor in ruling particular evidence admissible."

Reliable science for better justice

TrueAllele has been accepted 50 times as reliable evidence after admissibility challenge in US courts. In some cases, the court's main issue wasn't Daubert or reliability (e.g., timeliness). TrueAllele admissibility has been affirmed by appellate courts in eight states. This Third Circuit decision now adds a federal appellate court.

The take-home lesson from the Anderson affirmation is that TrueAllele PG software fully meets the Daubert reliability standard. Ineffective opposition arguments may be reworked for trial cross-examination, but they aren't relevant to the Daubert factors. For example, extraneous standards don't matter for determining forensic science reliability.

The Anderson and Gissantaner opinions are required reading for any forensic scientist or trial attorney preparing for a PG Daubert hearing. Against solid science, an opponent may try to suppress the facts, but courts usually accept well-presented reliable DNA evidence.

About the author

DNA Matters, an exclusive Forensic column, discusses cases that have been aided by the power of computer software in DNA analysis. It is authored by Dr. Mark Perlin, Ph.D., M.D., Ph.D., chief scientist, executive and founder at Cybergenetics. Twenty five years ago, Perlin invented TrueAllele® probabilistic genotyping for automated human identification from DNA mixtures. His company helped identify victim remains in the World Trade Center disaster, has worked for prosecution and defense in over a thousand cases, and has helped exonerate over 15 innocent men. He was a Scholar in Residence at Duquesne University’s Forensic Science and Law program, and is a Fellow of the American Academy of Forensic Sciences. © Mark Perlin 2026