Alternate-Perpetrator Evidence Requires a Concrete Nexus, and STRmix Discovery Must Materially Aid the Defense

Introduction

More than twenty years after Army Private First Class Amanda Gonzales was beaten and strangled in her barracks room on a U.S. military base in Germany, Shannon L. Wilkerson was prosecuted under 18 U.S.C. §§ 1111 and 3261(a)(2). Wilkerson, a married soldier who had been having an affair with Gonzales, was charged with first-degree murder. The jury instead convicted him of the lesser-included offense of second-degree murder.

The prosecution relied on motive, threats, opportunity, Wilkerson’s conduct on the night of the murder, his behavioral change afterward, and DNA recovered from a sweatshirt he wore that night. STRmix probabilistic-genotyping analysis indicated that the DNA mixture was substantially more likely under hypotheses including Gonzales, Wilkerson, and the sweatshirt’s owner than under competing hypotheses involving unknown contributors.

On appeal, Wilkerson challenged the exclusion of evidence suggesting another soldier was the killer, the denial of additional STRmix validation-study discovery, the denial of a new trial, the obstruction-of-justice enhancement, and the substantive reasonableness of his above-guidelines sentence.

Summary of the Opinion

The Eleventh Circuit affirmed on every issue:

• Alternate-perpetrator evidence: Evidence that Chinu Kim committed an unrelated double homicide five years later was too speculative. Kim’s proximity to Gonzales, his later ambiguous statement that he had “done this before,” and the partial covering of both victims did not establish a sufficient nexus to Gonzales’s murder.
• STRmix discovery: Wilkerson did not show that additional raw internal-validation data was material under Federal Rule of Criminal Procedure 16. The government had already produced the case-specific raw data, testing parameters, a validation report, and requested output files.
• DNA testimony: Any imprecision in describing likelihood ratios—such as saying “that” rather than “if”—did not constitute reversible plain error. The testimony as a whole made clear that the results expressed relative likelihoods rather than certainty.
• Weight of the evidence: This was not the exceptional case in which the evidence weighed so heavily against the verdict that allowing it to stand would cause a miscarriage of justice.
• Obstruction enhancement: The district court permissibly found that Wilkerson committed perjury when he denied killing Gonzales and claimed certainty despite professing gaps in his memory.
• Upward variance: The 360-month sentence was reasonable because the 2001 guideline range did not fully account for the killing of a pregnant fellow service member, the death of her unborn child, and the manner of the murder.

Analysis

1. Exclusion of Alternate-Perpetrator Evidence

Wilkerson sought to prove that Kim, who lived near Gonzales, might have killed her. In 2006, Kim murdered an auto mechanic and the mechanic’s son after a dispute over a car repair. Before those killings, Kim reportedly said that he had “done this before.” Wilkerson also emphasized that Kim concealed the mechanic’s body under a blanket, while Gonzales’s face had been covered by Army-issued pants.

The court held that these circumstances did not sufficiently connect Kim to the charged crime. Kim and Gonzales had no demonstrated relationship, Kim had no identified motive to kill her, and the later crimes involved shooting, kidnapping, stabbing, and mutilation rather than beating and strangulation. The fact that both crime scenes involved some form of covering was too generic to establish identity.

The ruling illustrates that the more relaxed treatment sometimes afforded defense evidence does not authorize speculation. Alternate-perpetrator evidence must connect the third party to the particular crime, not merely show that the third party is violent or later committed another murder.

2. Discovery of STRmix Validation Data

Rule 16 requires disclosure of scientific reports and data in the government’s possession when they are material to preparing the defense. Materiality requires more than an abstract relationship to the case; the requested information must meaningfully assist the defense or significantly alter its evidentiary position.

Wilkerson received the raw data from the testing of the sweatshirt, the testing parameters, USACIL’s 2017 validation report, and STRmix output information. His own experts replicated aspects of the analysis and challenged the government’s interpretation before the jury. One defense expert calculated a likelihood ratio of 79 for Gonzales, describing it as only limited support, while another argued that the mixture was complex, unbalanced, and potentially inconclusive.

Because Wilkerson could test the case-specific results and criticize the government’s methodology with the material already disclosed, the court concluded that his demand for all underlying internal-validation data was effectively a fishing expedition. Importantly, the opinion does not hold that raw validation data is never discoverable. It holds that this defendant did not demonstrate how the additional material would significantly change the proof in his favor.

3. DNA Likelihood Ratios and the New-Trial Motion

Wilkerson argued that the government’s expert improperly converted a conditional likelihood ratio into a statement that he and Gonzales were actual contributors. He also objected to the prosecutor’s isolated reference to a “1 in 4,000” ratio, which could be confused with a random-match probability.

Because no contemporaneous objection was made, plain-error review applied. The court found no controlling Supreme Court or Eleventh Circuit authority requiring experts to use a particular verbal formula when explaining likelihood ratios. Moreover, the testimony elsewhere used the appropriate conditional phrasing, and the defense experts extensively explained the limitations of the DNA evidence.

The prosecutor’s “1 in 4,000” question did not warrant relief because attorneys’ questions are not evidence, the jury was instructed accordingly, and the actual meaning of the 4,000 likelihood ratio was repeatedly stated correctly.

4. Weight of the Evidence

A new trial based on evidentiary weight is reserved for exceptional cases. The prosecution presented evidence that Wilkerson:

• had a motive arising from Gonzales’s pregnancy and threatened disclosure of the affair;
• said he would kill Gonzales if she reported him to command;
• was near her barracks and unaccounted for during the relevant period;
• asked other soldiers to provide him with a false alibi;
• later said, “I fucked up”;
• changed his behavior dramatically after the murder;
• had previously “play-choked” Gonzales; and
• wore a sweatshirt containing a DNA mixture consistent with Wilkerson, Gonzales, and its owner.

Although the defense offered competing interpretations of the DNA and challenged the timing and completeness of the investigation, those conflicts were for the jury. Wilkerson also testified, and the jury was entitled to reject his denials and consider them as substantive evidence of guilt.

5. Obstruction of Justice

U.S.S.G. § 3C1.1 permits a two-level enhancement when a defendant willfully obstructs justice, including through perjury. Perjury requires materially false testimony given willfully rather than through confusion, mistake, or faulty memory.

The district court independently found that Wilkerson’s categorical denial of the killing was false. Wilkerson argued that the jury might have believed he committed the crime while too intoxicated to remember it. The appellate court acknowledged that this was a possible view but held that the district court’s contrary interpretation—that the guilty verdict reflected rejection of Wilkerson’s denial—was also permissible. A choice between two permissible factual views is not clear error.

6. Upward Variance

Wilkerson’s advisory guideline range was 210 to 262 months, but the district court imposed 360 months. The court found that the 2001 Guidelines did not adequately reflect the killing of a pregnant fellow service member, the death of her unborn child, and the manner of the murder.

The sentencing judge noted that later Guidelines would produce a much higher range, but did not retroactively apply them. The later range served only as contextual support for the conclusion that the older range understated the offense’s seriousness. The district court also treated Wilkerson’s military service as mitigating but found that its weight was reduced by his adultery and the “ultimate betrayal” of killing another service member.

Finally, the court permissibly considered Wilkerson’s lack of remorse after he voluntarily allocuted. A defendant may maintain innocence, but once he speaks at sentencing, the court may evaluate what his statement reveals about remorse and acceptance of responsibility.

Precedents Cited

Evidence and Alternate-Perpetrator Liability

• United States v. Range established the district court’s broad discretion over evidentiary admissibility.
• United States v. Tinoco supplied the rule that reversal requires a clear abuse of discretion affecting substantial rights.
• United States v. Cohen recognized that admissibility standards are relaxed when a defendant offers evidence, while still prohibiting character evidence used merely to prove action in conformity with character.
• United States v. Starr required a nexus between an alleged alternate perpetrator and the particular charged crime; speculative or remote evidence may be excluded under Rule 403.

DNA Evidence and Discovery

• United States v. Gissantaner described STRmix as commonly used probabilistic-genotyping software that interprets mixed DNA samples through likelihood ratios.
• United States v. Jordan supplied the governing Rule 16 materiality standard: requested evidence must do more than bear an abstract relationship to the case and must be capable of significantly altering the defense’s quantum of proof.
• United States v. Cuya established abuse-of-discretion review for denials of motions to compel discovery.
• Dist. Attorney’s Off. for Third Jud. Dist. v. Osborne recognized the exceptional power of modern DNA evidence.
• McDaniel v. Brown emphasized that persuasive DNA evidence must be presented fairly and reliably.
• United States v. Martinez governed review of new-trial motions and stressed that weight-of-the-evidence relief is reserved for exceptional cases involving a serious risk of injustice.

False Testimony, Plain Error, and Evidentiary Weight

• Giglio v. United States formed the basis of Wilkerson’s claim that false or misleading DNA testimony required a new trial.
• United States v. Vallejo explained that a Giglio violation requires knowing use of perjured testimony, or failure to correct known false testimony, plus materiality.
• United States v. Moore supported plain-error review because Wilkerson did not object at trial.
• United States v. Steiger stated the four-part plain-error framework.
• United States v. De La Garza required a showing that the alleged error affected the outcome.
• United States v. Hesser held that there ordinarily is no plain error on an unresolved issue without controlling Supreme Court or Eleventh Circuit precedent and defined prejudice as a reasonable probability of a different result.
• United States v. Verdeza supported the conclusion that counsel’s questions are not evidence and that proper jury instructions may cure possible confusion.
• United States v. Brown allowed the jury to treat a defendant’s disbelieved testimony as substantive evidence of guilt.

Procedural Sentencing Review

• United States v. Carpenter recognized that an incorrectly calculated guideline range produces procedural unreasonableness.
• United States v. Cingari required de novo review of guideline interpretation and application.
• United States v. Isaac supplied clear-error review for factual findings and deferential review of guideline application to facts.
• United States v. Wall held that a factfinder’s choice between two permissible views of the evidence is not clearly erroneous.
• United States v. Stahlman defined perjury for § 3C1.1 purposes and required independent findings encompassing the factual predicates of perjury.

Substantive Sentencing Review

• United States v. Irey supplied the abuse-of-discretion standard and the requirement that appellate courts identify a clear sentencing error of judgment before vacating.
• United States v. Rosales-Bruno confirmed that sentencing courts may assign different weights to § 3553(a) factors and may vary based on disagreement with a guideline’s treatment of those factors.
• United States v. Butler cautioned appellate courts against second-guessing the weight assigned to proper sentencing considerations.
• United States v. Logal permitted consideration of a later, otherwise inapplicable guideline amendment when determining the degree of an upward departure or variance.
• United States v. Boone recognized that military service may carry both mitigating and aggravating significance.
• United States v. Stanley permitted a sentencing judge to consider the absence of remorse expressed in a defendant’s voluntary allocution.

Complex Concepts Simplified

Likelihood ratio

A comparison of two competing explanations for DNA evidence. A ratio of 4,000 means the DNA result is 4,000 times more likely under one stated contributor hypothesis than under the alternative hypothesis. It does not mean there is only a one-in-4,000 chance that the defendant is innocent.

Probabilistic genotyping

Computer-assisted analysis used to interpret DNA mixtures involving multiple people, especially when some contributors left only small amounts of DNA.

Rule 16 materiality

Discovery is material when it meaningfully helps prepare the defense. Merely showing that information relates generally to the scientific method is insufficient.

Alternate-perpetrator nexus

A concrete connection between another person and the charged crime, such as motive, opportunity, distinctive methods, physical evidence, or a relevant confession. General criminal propensity is not enough.

Plain error

A demanding standard applied when no timely objection was made. The defendant must establish an obvious error that affected the outcome and seriously undermined the proceeding’s fairness.

Procedural versus substantive reasonableness

Procedural review concerns how the sentence was calculated and explained. Substantive review concerns whether the final sentence is reasonable under the totality of the circumstances.

Upward variance

A sentence above the advisory guideline range based on the statutory sentencing factors. It differs from retroactively applying a later guideline.

Potential Impact

The opinion reinforces three practical limits on criminal-defense claims. First, later crimes by a proposed alternate perpetrator are not admissible merely because they show violent character; the defense must connect them to the charged offense through meaningful, case-specific similarities.

Second, defendants challenging probabilistic-genotyping evidence should identify precisely how requested validation data would alter expert analysis. Courts may deny broad demands where case-specific data and sufficient validation information have already enabled replication and cross-examination.

Third, sentencing courts retain broad discretion to conclude that an older guideline range understates the seriousness of an offense. Later guideline amendments may inform that assessment, provided the court does not retroactively substitute the amended range for the legally applicable one.

Because the opinion is unpublished, its formal precedential force is limited. Nevertheless, it offers useful guidance concerning STRmix discovery, alternate-perpetrator evidence, DNA terminology, perjury enhancements, and upward variances.

Conclusion

United States v. Shannon Wilkerson affirms that a defense theory must rest on more than speculation, scientific discovery must be materially useful rather than merely relevant in the abstract, and minor verbal imprecision in DNA testimony will not justify reversal without demonstrated prejudice. It also confirms substantial sentencing discretion where the advisory range fails to capture unusually aggravating circumstances. On the record presented, none of Wilkerson’s evidentiary, discovery, trial, or sentencing challenges established reversible error.