TrueAllele Probabilistic Genotyping Is Admissible Under Daubert Without Source-Code Disclosure (Third Circuit)

Introduction

In United States v. Hunter Ryan Anderson (3d Cir. Mar. 26, 2026), the Third Circuit addressed whether probabilistic genotyping software—specifically TrueAllele—satisfies the reliability requirements for expert testimony under Daubert v. Merrell Dow Pharmaceuticals, Inc. and Federal Rule of Evidence 702.

The underlying prosecution arose from a search warrant execution in which police seized a firearm from a bag that also contained Anderson’s identification and magazines. The Pennsylvania State Police Crime Laboratory identified a multi-source DNA mixture from the gun swab but could not declare a match “within a degree of scientific certainty.” Law enforcement then sent the resulting profiles to a private company (Cybergenetics, Corp.), which used TrueAllele to compute a likelihood ratio of 11.5 trillion supporting that Anderson contributed DNA to the mixture.

Anderson moved to exclude the TrueAllele evidence under Daubert and separately challenged 18 U.S.C. § 922(g)(1) under the Second Amendment. After an adverse ruling, he pleaded guilty while preserving those issues, and he also challenged his sentence (including consecutive treatment relative to expected state parole-violation proceedings).

Summary of the Opinion

The Third Circuit affirmed across the board. Most significantly, it held that the government met its burden to show TrueAllele’s probabilistic genotyping methodology was “reliable enough” to be admissible under Rule 702. The Court emphasized that alleged weaknesses—such as operator choices, the probabilistic nature of results, or speculative software-bug concerns—are primarily for cross-examination, not categorical exclusion.

The Court also rejected Anderson’s Second Amendment challenges as foreclosed by circuit precedent because he was on parole when arrested, and it found no sentencing abuse of discretion in imposing a consecutive sentence.

Analysis

Precedents Cited

1) The governing admissibility framework (Daubert / Rule 702 gatekeeping)

  • Daubert v. Merrell Dow Pharmaceuticals, Inc.: The foundational reliability framework. The Court treated Daubert’s factors (testability, peer review, error rates, standards, general acceptance) as flexible “guideposts,” not a checklist.
  • In re Paoli R.R. Yard PCB Litig.: The Court leaned on Paoli for the purpose and mechanics of the Daubert hearing, including that admissibility must be shown by a preponderance under Rule 104(a), and that non-judicial uses are relevant to assessing reliability.
  • Elcock v. Kmart Corp.: Provided the tripartite Rule 702 structure—qualifications, reliability, and fit—while confirming reliability was the only real issue in dispute.
  • In re TMI Litig.: Supplied the key admonition that Rule 702 reliability is a lower bar than “correctness” on the merits and supported the role of peer review as a “component of good science.”
  • United States v. Mitchell: The most important Third Circuit analogue in the opinion. Mitchell supplied (i) the flexible treatment of Daubert factors, (ii) the Court’s focus on false positives for error-rate discussions, and (iii) the principle that human judgment and “subjective, sliding-scale” elements do not automatically defeat reliability. Mitchell also reinforced that testability turns on whether a method can be challenged objectively and meaningfully through adversarial testing.
  • Cohen v. Cohen: Cited for the abuse-of-discretion standard on Daubert rulings and the “rigorous gatekeeping” function.

2) Third Circuit and sister-circuit DNA/software comparators

  • United States v. Trala: A key in-circuit stepping stone: the Third Circuit previously upheld reliability of a DNA method using software and statistics, though “not as complex as TrueAllele.” Anderson frames its holding as a “natural extension” of Trala.
  • United States v. Gissantaner (6th Cir.): Upheld STRmix probabilistic genotyping under Daubert. The Third Circuit expressly agreed with “substantially all” of Chief Judge Sutton’s points and used Gissantaner to explain likelihood ratios and lab testing paradigms for mixture interpretation.

3) Fairness, disclosure, and adversarial testing (without turning Daubert into discovery)

  • Dennis v. Sec'y, Pa. Dep't of Corr. and United States v. Scarfo: Cited to underscore constitutional disclosure duties (Brady/Giglio-type principles) concerning favorable evidence and credibility impeachment, positioning those doctrines—not Daubert—as primary fairness tools.
  • United States v. Ellis: Noted as a district court matter in which source code review was permitted under a protective order, but it did not produce appellate guidance and did not support blanket entitlement to source code in this case.

4) State-court decisions on source code and probabilistic genotyping admissibility

  • People v. Wakefield, State v. Simmer, and Commonwealth v. Foley: Cited as aligned authority rejecting the notion that source-code access is required to establish admissibility (even though those courts were applying non-Daubert standards).

5) Broader forensic/DNA context and “probabilistic” expert opinions

  • Dist. Att'y's Off. for Third Jud. Dist. v. Osborne and Maryland v. King: Used to situate DNA as powerful but not absolute proof; the Court stresses analytic caution with forensic DNA.
  • United States v. Mornan: Cited to normalize expert opinions expressed probabilistically rather than with certainty.

6) General acceptance and defining the “relevant scientific community”

  • United States v. Velasquez: Supported the proposition that even disputes about measurable standards do not necessarily bar expert evidence.
  • Kannankeril v. Terminix Int'l, Inc.: Reinforced that peer review/publication is helpful but not a necessary condition of reliability.

7) The opinion’s “non-binding authority” discussion

  • United States v. Russell (9th Cir.): Distinguished as non-precedential, record-thin (no Daubert hearing), and reliant on a draft NIST report language later deleted in the final version. The Third Circuit treated Russell as unpersuasive on this record.

8) The other issues affirmed

  • United States v. Moses: Foreclosed the as-applied and facial Second Amendment challenges to § 922(g)(1) on these facts because Anderson was on parole.
  • Setser v. United States: Supported the district court’s discretion to run the federal sentence consecutive to anticipated state parole-related sentencing.
  • United States v. Harris and United States v. Jumper: Provided review standards for the Second Amendment issue (de novo) and sentencing (abuse of discretion).

Legal Reasoning

1) The “reliable enough” theme and the role of cross-examination

The Court repeatedly returned to a central Daubert/Rule 702 premise: admissibility is not a guarantee of correctness. Drawing from In re TMI Litig., it framed reliability as a threshold question. Once a method has adequate scientific grounding, remaining disputes largely go to weight, to be tested through cross-examination and competing experts.

2) Applying the Daubert guideposts to TrueAllele

The district court conducted a two-day Daubert hearing with expert testimony on both sides. On appeal, the Third Circuit concluded the government carried the Rule 104(a) preponderance burden on each major reliability indicator:

  • Testability: TrueAllele can be tested by constructing known DNA mixtures and checking for false negatives and false positives. The defense also had meaningful opportunities to run software-based testing (access to the algorithm, data, and a TrueAllele machine environment; disclosure of comparisons to other probabilistic programs; and materials about the server and VUIer components).
  • Source code not required for Daubert testability: The Court rejected the defense contention that lack of source code access defeated testability. It reasoned that (i) Daubert is about whether a method can be objectively challenged, (ii) the defense already had access to the algorithm (the scientific logic), and (iii) discovery fairness is governed by other tools (Rule 16, expert-disclosure rules, Brady/Giglio principles).
  • Error rates: Emphasizing United States v. Mitchell’s focus on false positives, the Court credited the government’s evidence (including a 2014 study reporting a 0.005% false positive rate and testimony that for very large likelihood ratios the false positive probability becomes “infinitesimally small,” with an estimate of ~1 in 146 trillion in this case). Speculation about generic coding error rates and hypothetical bugs was treated as insufficient to undermine reliability absent case-specific linkage.
  • Standards: The Court accepted that probabilistic genotyping is governed by standards (e.g., 2015 validation guidelines and the 2020 ANSI/AAFS standard). It rejected an attack premised on the state lab’s supposed lack of calibration to TrueAllele because the lab did not perform probabilistic genotyping; it performed “wet lab” steps not challenged in the motion.
  • Peer review/publication: The record reflected peer-reviewed validation studies. The Court rejected the suggestion that authorship overlap (the government expert helped author several) defeated the factor, especially given anonymous independent peer review.
  • General acceptance: The Court credited evidence of broad use for decades, by hundreds of organizations, including criminal defense teams, and multiple admissibility rulings. It also clarified that the relevant community is forensic DNA interpretation—not “computer scientists and software engineers”—echoing how United States v. Mitchell and United States v. Trala framed the “community” around the forensic identification discipline.

3) Treatment of NIST-related criticisms raised via Russell

The Court’s handling of United States v. Russell is notable: rather than treating NIST commentary as a dispositive external veto, the Court kept the focus on Daubert’s traditional touchstones—peer review, testing, error rates, and acceptance—applied to a developed evidentiary record. It also discounted reliance on draft-report language that the NIST later deleted in its final report.

Impact

  • Third Circuit endorsement of probabilistic genotyping (TrueAllele): The decision positions TrueAllele as generally admissible under Rule 702 in federal trials within the circuit when supported by a record like the one developed here—likely reducing successful blanket challenges to probabilistic genotyping as a category.
  • Limits on “source code” demands framed as Daubert requirements: A key practical precedent is the Court’s refusal to transform Daubert into a vehicle for compelled source-code inspection. Defendants can still litigate disclosure under criminal discovery doctrines, but Anderson signals that source code is not inherently required to establish “testability.”
  • Clarification of the “relevant scientific community”: By resisting efforts to redefine the community to include software engineering, the Court reduces a common defense strategy: demanding acceptance across adjacent technical fields rather than the forensic field actually using the method.
  • Future hearings will likely pivot from categorical challenges to case-specific ones: The opinion invites more granular litigation about operator choices, mixture complexity, and parameter selection as weight/credibility issues—i.e., cross-examination topics—rather than admissibility blockers.

Complex Concepts Simplified

Probabilistic genotyping
A statistical method (often software-driven) for interpreting complex DNA mixtures by modeling different contributor scenarios and calculating how strongly the data supports one scenario over another.
DNA mixture (multi-source mixture)
A sample containing DNA from multiple people. Mixtures complicate interpretation because more than two alleles may appear at a locus and contributors may contribute uneven amounts of DNA.
Locus / allele / short tandem repeat
A locus is a standardized location on a chromosome analyzed in forensic testing. An allele is the genetic variant at that locus (typically two per person, one from each parent). Short tandem repeats are repeating DNA sequences measured to distinguish individuals.
Likelihood ratio
A comparison of probabilities: how likely the observed DNA data is if the defendant contributed to the mixture versus how likely it is if a random unrelated person contributed. A very large number (like 11.5 trillion) strongly favors the “defendant contributed” hypothesis.
False positive vs. false negative (in this context)
A false positive would mean the software indicates a person contributed DNA when they did not. A false negative would mean it fails to detect a real contributor in a test scenario.
Daubert “reliability” vs. “correctness”
Reliability asks whether the method rests on a sound scientific footing and can be responsibly used; it does not require the court to conclude the method’s result is definitively correct in the case.

Conclusion

United States v. Hunter Ryan Anderson makes two practical contributions to federal evidence law in the Third Circuit: (1) it endorses TrueAllele probabilistic genotyping as sufficiently reliable under Daubert and Rule 702 on a robust record, and (2) it rejects the notion that source-code disclosure is a prerequisite to establishing testability and admissibility. The opinion reinforces a familiar Daubert boundary: courts screen for methodological reliability, while juries (aided by cross-examination and competing experts) resolve residual disputes about imperfections, assumptions, and weight.