I respectfully dissent. I disagree with the majority’s conclusion that the evidence supports a finding of constructive possession. That conclusion makes it necessary to address the prosecution’s alternate theory of actual possession. As regards evidence of actual possession, I would also conclude that the district court violated Paul Reyes Sedillo’s confrontation rights by admitting the testimony of Carrie Zais, a DNA expert, and that this error was not harmless beyond a reasonable doubt. Accordingly, I would reverse and remand with directions to vacate Sedillo’s conviction of being a felon in possession of a firearm.
I
The majority concludes that the evidence presented at trial was sufficient to establish that Sedillo had constructive possession of the firearm. Sedillo resided in his mother’s home with his mother, Lá-veme Webb. In cases of joint occupancy, as is the case here, constructive possession requires that there be “some connection or nexus between the defendant and the firearm.” United States v. King, 632 F.3d 646, 651 (10th Cir.2011) (quoting United States v. Ledford, 443 F.3d 702, 713 (10th Cir.2005)). To meet the nexus requirement, the government must “point to evidence plausibly supporting the inference that the defendant had knowledge of and access to the firearm.” United States v. Poe, 556 F.3d 1113, 1125 (10th Cir.2009). The evidence here does not establish a sufficient nexus between Sedillo and the firearm.
The majority points out that “Mr. Sedil-lo does not deny that he lived in the house, nor does he suggest that he could not access the closet.” Maj. Op. 8. However, these facts do not establish that Sedillo had knowledge of the firearm. The majority concludes that “in the case of two occupants where one occupant denies knowledge of a shotgun, it is plausible to infer the other occupant knew about the shotgun.” Maj. Op. 8. However, this assertion is unsupported by our case law and contravenes our precedent. In United States v. Jones, 49 F.3d 628 (10th Cir.1995), we explained that it is improper to “pile[ ] one inference upon another” in order to establish possession. Id. at 633. The majority here engages in precisely this improper piling of inferences when it finds it “plausible to infer” that Sedillo had knowledge of the shotgun from his mother’s lack of knowledge of the shotgun.
The cases that the majority cites reaffirm the knowledge requirement. In United States v. King, 632 F.3d 646 (10th Cir.2011), we held that the knowledge requirement was satisfied because the defendant claimed ownership of the firearm twice in front of law enforcement officers, and because a photograph of the firearm was found on the defendant’s cell phone. Id. at 654. Similarly, in United States v. Mendez, 514 F.3d 1035 (10th Cir.2008), we determined that the defendant had knowledge of the firearms because the defendant admitted to knowing that the firearms were in the house. Id. at 1042. See also Poe, 556 F.3d at 1125 (finding that knowledge of the weapon was established when the defendant told the police “[t]he dope and the gun are mine”). Here, there is no evidence that Sedillo had knowledge of the shotgun.
The majority’s use of “inference” evidence becomes even more shaky when we acknowledge that there were other individ*682uals, aside from Sedillo and his mother, who had access to the house. This evidence makes it just as “plausible to infer” that someone else brought the shotgun into the house. According to Webb’s testimony, her sister also had a key to the house. Ill R. 477. And significantly, on the night of the incident (October 22, 2009), Sedillo’s brother, Roman Reyes, slept in a car parked in front of the house. In her testimony, Webb acknowledged that it was possible that Roman Reyes came into the house on October 22, 2009, and talked with Sedillo at some point. Id. at 480. Roman Reyes was living in Webb’s house when Sedillo first moved into the house in August of 2009. On September 16, 2009, Roman Reyes and Sedillo had an argument and Roman Reyes left Webb’s house to stay with other relatives. Id. Although Webb testified that Roman Reyes did not have a key to the house, Roman Reyes testified that he still had a key to Webb’s house and that Webb knew that he still had a key. Id. at 475, 512. When asked at trial whether he kept a shotgun in Webb’s house, Roman Reyes refused to answer the question without an attorney present because he also has a prior felony conviction. Id. at 512.
The majority dismisses the evidence that someone else might have brought the firearm into the house. The majority notes that “defense counsel, in its closing argument, suggested three times that Mr. Reyes, Mr. Sedillo’s brother, was responsible for the shotgun.” Maj. Op. 680. The majority concludes that “[t]he jury was free to consider this point in reaching its decision,” and that “the jury found this argument unpersuasive.” Maj. Op. 680. That the jury found Sedillo guilty does not relieve us, or the district court, from requiring an evidentiary basis for the verdict which satisfies the legal requirements for constructive possession.1 As noted in the discussion that follows, when the legal requirements for constructive possession are not met, this court will reverse the conviction at issue regardless of the jury’s guilty verdict.
Further, I find unpersuasive the majority’s distinction of this case from our prior decisions. In United States v. Mills, 29 F.3d 545 (10th Cir.1994), we found insufficient evidence for a jury to find constructive possession when the defendant’s housemate testified that she placed firearms in the dining room table compartment without the defendant’s knowledge and contrary to his instructions. Id. at 550. But we also noted in Mills that even if the jury disbelieved the house mate’s testimony, “that disbelief cannot constitute evidence of the crimes charged and somehow substitute for knowing constructive possession in this joint occupancy situation.” Id. As we explained, the defendant was not required to prove that he was denied access to the kitchen table or compartment, but that “the government had to come forward with evidence to connect [the defendant] with knowing constructive *683possession of the firearms.” Id. As is the case here, “[m]ere dominion or control ... is insufficient to establish constructive possession.” Id.
And in United States v. Hishaw, 235 F.3d 565 (10th Cir.2000), we determined that testimony at trial regarding the defendant’s prior possession of the firearm was “simply too remote and too vague to support the inference that [defendant] constructively possessed the pistol.” Id. at 572. Likewise, in United States v. Taylor, 113 F.3d 1136 (10th Cir.1997), we held that witness testimony of the defendant’s possession of “a small handgun on one or two occasions” was insufficient to establish that the defendant “ever possessed or even had knowledge of’ the firearm that was discovered in the defendant’s apartment. Id. at 1145. Additionally, in Taylor, we found that the evidence was insufficient to support a constructive possession conviction when the firearm was found in the defendant’s bedroom. We noted that the defendant shared the bedroom with another occupant, and even though the defendant’s belongings were found in the bedroom, we held that “joint occupancy of a bedroom, without more, is insufficient to support a conviction of constructive possession of a gun found in a bedroom.” Id. at 1146 (citing United States v. Sullivan, 919 F.2d 1403, 1431 (10th Cir.1990)). Here, the firearm was not found in Sedillo’s bedroom, but in a hall closet in a common area of the house.
The majority distinguishes Taylor from this case because there were three occupants in the apartment in Taylor, whereas there were only two known occupants in this case. The number of joint occupants is not sufficient to distinguish Taylor from this case. “In cases of joint occupancy, where the government seeks to prove constructive possession by circumstantial evidence, it must present evidence to show some connection or nexus between the defendant and the firearm or other contraband.” Id. at 1145 (quoting Mills, 29 F.3d at 549). There is no evidence showing a connection or nexus between Sedillo and the shotgun.
For these reasons, I would conclude that the evidence is insufficient to uphold Sedil-lo’s conviction which rests upon a theory of constructive possession.
II
Because I would conclude that the evidence does not support a finding of constructive possession, I must next address whether there was sufficient evidence to support a finding of actual possession. This court reviews district court’s eviden-tiary rulings for an abuse of discretion. United States v. Ledford, 443 F.3d 702, 707 (10th Cir.2005); United States v. Jenkins, 313 F.3d 549, 559 (10th Cir.2002). If error is found and a party objects to the court’s evidentiary ruling based solely on the Federal Rules of Evidence, this court reviews for nonconstitutional harmless error. Ledford, 443 F.3d at 707. However, if error is found and a party objects to evidentiary rulings that implicate the Confrontation Clause, this court “may find error harmless only if it is convinced that the error was harmless beyond a reasonable doubt.” Id. (citing United States v. Jefferson, 925 F.2d 1242, 1253-54 (10th Cir.1991)).
The government’s proof of actual possession rested upon the testimony of the DNA expert, Carrie Zais. Sedillo objected to Zais’s testimony that the DNA found on the shotgun came from Sedillo on grounds that her testimony would violate his confrontation rights. I R. 68-73. Accordingly, any violations of Sedillo’s Confrontation Clause rights are subject to harmless error review, in which this court determines whether the error was harmless beyond a reasonable doubt. This court reviews the *684record de novo. United States v. Chavez, 481 F.3d 1274, 1277 (10th Cir.2007).
I would conclude that Zais’s testimony violated the Confrontation Clause of the Sixth Amendment. If Zais’s testimony is excluded, there is insufficient evidence to support a conviction of actual possession. And, as a result, any error in admitting Zais’s testimony would not be harmless beyond a reasonable doubt because Sedil-lo’s conviction cannot be sustained based on a theory of constructive possession.
The Confrontation Clause of the Sixth Amendment provides that “[i]n all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him.” U.S. Const. amend. VI. In Crawford v. Washington, 541 U.S. 36, 59, 124 S.Ct. 1354, 158 L.Ed.2d 177 (2004), the Supreme Court interpreted the Confrontation Clause to mean that “[testimonial statements of witnesses absent from trial [can be] admitted only where the declarant is unavailable, and only where the defendant has had a prior opportunity to cross-examine.” Statements are testimonial when “the primary purpose of the interrogation is to establish or prove past events potentially relevant to later criminal prosecution.” Davis v. Washington, 547 U.S. 813, 822, 126 S.Ct. 2266, 165 L.Ed.2d 224 (2006).
The Supreme Court held in Melendez-Diaz v. Massachusetts, 557 U.S. 305, 129 S.Ct. 2527, 174 L.Ed.2d 314 (2009), and Bullcoming v. New Mexico, — U.S. -, 131 S.Ct. 2705, 180 L.Ed.2d 610 (2011), that scientific reports could not be used as substantive evidence against the defendant at trial unless the analyst who had prepared the report was subject to cross-examination. In Melendez-Diaz, the defendant was charged with distributing and trafficking cocaine. At trial, the prosecution submitted three “certificates of analysis” showing that forensic analysis of seized substances from the defendant found the substances to contain cocaine. Melendez-Diaz, 557 U.S. at 308, 129 S.Ct. 2527. The Court held that the certificates of analysis, which were sworn under oath and notarized, were “functionally identical to live, in-court testimony, doing ‘precisely what a witness does on direct examination.’ ” Id. at 310-11, 129 S.Ct. 2527 (quoting Davis, 547 U.S. at 830, 126 S.Ct. 2266). As a result, the defendant was “entitled to be confronted with the analysts at trial.” Id. at 311, 129 S.Ct. 2527 (quotation omitted). In Bullcoming, the defendant was convicted of driving under the influence of intoxicating liquor. At trial, the court admitted into evidence a forensic laboratory report certifying that the defendant’s blood alcohol concentration was above the legal threshold. Instead of calling the analyst who signed the certification to testify, the prosecution called another analyst who was familiar with the laboratory’s general testing procedures but who had not participated nor observed the actual analysis on the defendant’s blood. The Court rejected this surrogate testimony and held that “[t]he accused’s right is to be confronted with the analyst who made the certification.” Bullcoming, 131 S.Ct. at 2710.
In her concurrence to Bullcoming, Justice Sotomayor highlighted several factual circumstances that Bullcoming did not present. First, Justice Sotomayor explained that Bullcoming “is not a case in which the State suggested an alternate purpose ... for the [forensic] report.” Id. at 2722 (Sotomayor, J., concurring). Second, Bullcoming “is not a case in which the person testifying is a supervisor, reviewer, or someone else with a personal, albeit limited, connection to the scientific test at issue.” Id. Third, Justice Sotomayor highlighted that Bullcoming “is not a case in which an expert witness was asked for his independent opinion about underlying *685testimonial reports that were not themselves admitted into evidence.” Id. Finally, Justice Sotomayor noted that Bullcom-ing “is not a case in which the State introduced only machine-generated results.” Id. According to Justice Sotoma-yor, Bullcoming “does not present, and thus the Court’s opinion does not address, any of these factual scenarios.” Id.
In Williams v. Illinois, — U.S. -, 132 S.Ct. 2221, 183 L.Ed.2d 89 (2012), the most recent Supreme Court case to address the admission of scientific reports under the Confrontation Clause, the Court sought to respond to one of Justice Soto-mayor’s factual distinctions, the question of “determining] the constitutionality of allowing an expert witness to discuss others’ testimonial statements if the testimonial statements were not themselves admitted as evidence.” Id. at 2233 (quoting Bullcoming, 131 S.Ct. at 2722 (Sotomayor, J., concurring)). The defendant in Williams was convicted of rape. At trial, the prosecution called an expert witness who testified that a DNA profile produced by an outside laboratory, Cellmark, matched a profile produced by the state police lab using a sample of the defendant’s blood. Id. at 2227. The defendant in Williams argued that the Confrontation Clause was violated when the witness “referred to the DNA profile provided by Cellmark as having been produced from semen found on the victim’s vaginal swabs.” Id. According to the defendant, his right to confrontation was violated because the witness did not have personal knowledge that the profile produced by Cellmark was based on vaginal swabs taken from the victim. The Supreme Court in Williams split three ways: Justice Alito wrote a plurality that three other Justices joined; Justice Breyer (who joined in the plurality) and Justice Thomas (who did not join in the plurality) each wrote a concurring opinion; and Justice Kagan, with whom three other Justices joined, wrote for the dissent.
The plurality concluded that the testimony at issue in Williams did not violate the Confrontation Clause. The plurality rested on two independent bases in reaching its decision. First, the plurality found that the Confrontation Clause does not apply to “out-of-court statements that are not offered to prove the truth of the matter asserted.” Id. at 2228. The plurality reasoned that the phrase “[DNA] found in semen from the vaginal swabs of [the victim]” was not offered to prove the truth of the matter asserted; instead, “that fact was a mere premise of the prosecutor’s question.” Id. at 2236. Because Williams was a bench trial instead of a jury trial, the plurality assumed that the trial judge understood that the phrase was not admissible to prove the truth of the matter asserted. Id. at 2236-37. On a second, independent basis, the plurality concluded that even if the Cellmark forensic report had been admitted into evidence and introduced for its truth, the Confrontation Clause was not violated because the report was not testimonial. The plurality found that the primary purpose of the report was not to accuse the defendant or to create evidence for use at trial. Instead, “its primary purpose was to catch a dangerous rapist who was still at large, not to obtain evidence for use against petitioner, who was neither in custody nor under suspicion at that time.” Id. at 2243-44 (emphasis added).
Justice Thomas had a separate reason for concluding that the expert witness’s testimony did not violate the Confrontation Clause. Justice Thomas in his concurrence agreed with the dissent that the expert witness’s statements regarding the Cellmark report were introduced for their truth, and that the plurality’s “primary purpose” test in determining whether a *686statement is testimonial is unjustified. Id. at 2258-59, 2261-64 (Thomas, J., concurring in judgment). However, he concluded that the Cellmark report was not testimonial because the report “lacked the requisite ‘formality and solemnity’ to be considered ‘testimonial’ for purposes of the Confrontation Clause.” Id. at 2255 (citing Michigan v. Bryant, — U.S. -, 131 S.Ct. 1143, 1167, 179 L.Ed.2d 93 (2011) (Thomas, J., concurring in judgment)). According to Justice Thomas, the Cellmark report lacked “the solemnity of an affidavit or deposition” because it was not a sworn or certified declaration of fact, nor was it “the product of any sort of formalized dialogue resembling custodial interrogation.” Id. at 2260.
The dissent in Williams disagreed with both reasons offered by the plurality, as well as Justice Thomas’s “indicia of solemnity” test. The dissent argued that the Cellmark report in Williams is identical to the one in Bullcoming and Melendez-Diaz, and so “the substance of the report could come into evidence only if Williams had a chance to cross-examine the responsible analyst.” Id. at 2266-67 (Kagan, J., dissenting).
The Williams opinion is no doubt a “fractured decision.” Id. at 2265 (Kagan, J., dissenting). None of the three rationales proffered by the plurality, Justice Thomas’s concurrence, or the dissent garners the support of a majority of the Court. As Justice Kagan noted, five Justices disagreed with the reasoning offered by the plurality:
I call Justice ALITO’s opinion “the plurality,” because that is the conventional term for it. But in all except its disposition, his opinion is a dissent: Five Justices specifically reject every aspect of its reasoning and every paragraph of its explication.
Id.
In Marks v. United States, 430 U.S. 188, 97 S.Ct. 990, 51 L.Ed.2d 260 (1977), the Supreme Court held that “[wjhen a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, ‘the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds.’ ” Id. at 193, 97 S.Ct. 990 (quoting Gregg v. Georgia, 428 U.S. 153, 169 n. 15, 96 S.Ct. 2909, 49 L.Ed.2d 859 (1976)). “In practice, however, the Marks rule produces a determinate holding ‘only when one opinion is a logical subset of other, broader opinions.’ ” United States v. Carrizales-Toledo, 454 F.3d 1142, 1151 (10th Cir.2006) (quoting King v. Palmer, 950 F.2d 771, 781 (D.C.Cir.1991) (en banc)). Applying Marks becomes problematic “[wjhen the plurality and concurring opinions take distinct approaches, and there is no ‘narrowest opinion’ representing the ‘common denominator of the Court’s reasoning.’ ” Id. (quoting King, 950 F.2d at 781). Consequently, this court “do[esj not apply Marks when the various opinions supporting the Court’s decision are mutually exclusive.” Id. See also Large v. Fremont Cnty., Wyo., 670 F.3d 1133, 1141 (10th Cir.2012).
Based on the facts presented, I need not determine the precise holding of Williams to conclude that Zais’s testimony violated the Confrontation Clause because her statements would violate Sedillo’s confrontation rights under both lines of reasoning offered by the plurality.2
*687 Whether the Statement was Offered for the Truth of the Matter Asserted
In Williams, the plurality’s first independent basis for concluding that the statement at issue did not violate the Confrontation Clause is that the statement was not offered for the truth of the matter asserted. According to the plurality, “the Confrontation Clause ‘does not bar the use of testimonial statements for purposes other than establishing the truth of the matter asserted.’ ” Williams, 132 S.Ct. at 2235 (quoting Crawford, 541 U.S. at 59-60, n. 9, 124 S.Ct. 1354). And under the rules of evidence, “an expert may base an opinion on facts that are ‘made known to the expert at or before the hearing,’ but such reliance does not constitute admissible evidence of this underlying information.” Id. (quoting Ill. R. Evid. 703; Fed.R.Evid. 703). Accordingly, the plurality concluded that the testimony at issue in Williams— the phrase “found in semen from the vaginal swabs of [the victim]” — was “not admissible for the purpose of proving the truth of the matter asserted.” Id. at 2236. Instead, the statement was “a mere premise of the prosecutor’s question” that the expert witness “assumed ... to be true when she gave her answer indicating that there was a match between the two DNA profiles.” Id.
The statements at issue in this case are Zais’s disclosure, through her expert testimony, of the laboratory report that was prepared by Stephanie Willard. Specifically, the statements at issue are Zais’s opinion that the DNA profile derived from evidence recovered from the shotgun matched Sedillo’s profile, Zais’s reference to contents of the forensic laboratory report and Willard’s notes, and Zais’s explanation of Willard’s testing procedures. According to Sedillo, Zais’s opinion that the DNA profile recovered from the shotgun matched Sedillo’s profile “necessarily assumes that both [DNA] profiles are accurate,” and that her opinion “is in no way independent of that assumption.” Aplt. Open. Br. 14.
There are several factual distinctions between Williams and this case. The plurality in Williams pointed out four safeguards to prevent the disclosure of out-of-court statements on which the expert relied: (1) trial courts can screen out experts who act as mere conduits for hearsay by enforcing the requirements for expert testimony, (2) experts are generally precluded from disclosing inadmissible evidence to a jury, (3) if such evidence is disclosed to the jury, the trial judge may and, under most circumstances, must, instruct the jury that out-of-court statements cannot be accepted for their truth, and (4) if the prosecution cannot provide any indepen*688dent admissible evidence to prove the foundational facts that are essential to the relevance of the expert’s testimony, then the expert’s testimony cannot be given any weight by the trier of fact. Williams, 132 S.Ct. at 2241. Several of these safeguards are not present in this case, suggesting that Zais’s testimony would violate the Confrontation Clause under the plurality’s first line of reasoning.
First, the plurality in Williams drew a distinction between a bench trial and a jury trial when addressing the disclosure of an expert witness’s underlying factual assumptions. While the rules of evidence generally bar the disclosure of such inadmissible evidence in jury trials, there are “no restriction[s] on the revelation of such information to the factfinder” in bench trials because “it is presumed that the judge will understand the limited reason for the disclosure of the underlying inadmissible information and will not rely on that information for any improper purpose.” Williams, 182 S.Ct. at 2234-35. The plurality further explained that the testimony at issue in Williams could not have gone to the jury in a jury trial “[ajbsent an evaluation of the risk of juror confusion and careful jury instructions.” Id. at 2236.
Unlike Williams, which involved a bench trial, the case at bar was tried to a jury. Significantly, there was no evaluation of the risk of juror confusion regarding the underlying assumption of Zais’s testimony — that the DNA profile generated from evidence on the shotgun and Sedillo’s DNA profile were accurate. The plurality in Williams explained that “the trial judges may and, under most circumstances, must, instruct the jury that out-of-court statements cannot be accepted for their truth, and that an expert’s opinion is only as good as the independent evidence that establishes its underlying premises.” Id. at 2241. The district court here did not give the jury this limiting instruction. Instead, the district court gave this instruction regarding expert testimony: “Give opinion testimony as much weight as you think it deserves, considering the education and experience of the witness, the soundness of the reasons given for the opinion, and other evidence in the trial.” I R. 123. Under the Federal Rules of Evidence, an expert witness may base an opinion on facts or data that are otherwise inadmissible, but “the proponent of the opinion may disclose them to the jury only if their probative value in helping the jury evaluate the opinion substantially outweighs their prejudicial effect.” Fed.R.Evid. 703.
In the present case, Willard analyzed the DNA profiles and prepared the forensic laboratory report. However, it was Zais who testified at trial that Sedillo’s DNA profile matched the DNA profiled recovered from the shotgun. The district court relied on this court’s initial ruling in Pablo before remand and found that “an expert testifying about another DNA analyst’s work is proper, ‘so long as the expert opinion is that of the expert appearing at trial,’ rather than someone simply reciting what the first analyst did.” Aplt. Open. Br. Attach. B at 2 (citing and quoting Pablo, 625 F.3d at 1292-96). The district court found no Confrontation Clause problem because Zais can testify to her opinion, based on Willard’s notes and test results, that the DNA profiles matched. Id. at 3. However, the district court did not determine whether the probative value of Willard’s underlying analysis substantially outweighs its prejudicial effect.
Second, the plurality in Williams also emphasized that the expert witness in that case testified to the truth of matters for which she had personal knowledge, but did not testify to the truth of the underlying laboratory report:
*689[The expert witness] made no other reference to the Cellmark report, which was not admitted into evidence and was not seen by the trier of fact. Nor did she testify to anything that was done at the Cellmark lab, and she did not vouch for the quality of Cellmark’s work.
Williams, 132 S.Ct. at 2235. In the present case, the report that Willard prepared was not admitted into evidence, but Zais did refer to the report during her testimony, as well as how Willard performed her testing and her analysis. Zais explained how Willard swabbed the textured parts of the stock, lever, safety, and trigger of the gun for DNA. Ill R. 242. Zais then described the quantitation process, a test to determine how much DNA is present, of the sample from the shotgun. Id. at 244. Zais reported that through the quantitation process, Willard determined that the sample was dirty. Id. at 245. Zais then testified that Willard diluted the sample with water so that the sample could be tested, and she testified to the amount of DNA recovered from the shotgun and control samples collected from other parts of the gun. Id. at 244, 246-47. As technical reviewer, Zais did not participate or observe the analysis that Willard conducted, but she reviewed Willard’s case notes, Willard’s conclusions in the report, and agreed with Willard’s conclusions. Id. at 241-42. Unlike Williams, where the expert witness did not make references to the laboratory report or testify to anything that was done at the laboratory, here, Zais testified as to Willard’s testing procedure, referred to Willard’s report, and vouched for the quality of Willard’s work.
Further, the plurality in Williams cited “strong circumstantial evidence regarding the reliability of Cellmark’s work.” Williams, 132 S.Ct. at 2239. The plurality questioned “how ... shoddy or dishonest work in the Cellmark lab [could] have resulted in the production of a DNA profile that just so happened to match petitioner’s.” Id. (footnote omitted). Under the facts of Williams, where Williams was not even under suspicion when Cellmark profiled the semen sample, it was permissible for the trier of fact to infer that the odds of sloppy work or a dishonest lab technician to falsely accuse the petitioner was “exceedingly low.” Id. Compared to Williams, there is a lack of such “strong circumstantial evidence” in this case. Sed-illo was already identified as a suspect when DNA sample was collected from the shotgun. DNA samples from Sedillo were sent to the Scientific Laboratory Division of the New Mexico Department of Public Safety along with the shotgun. Willard was aware of Sedillo’s identity when she performed the testing, and Willard knew that this was a felon in possession case before she conducted the DNA test. Id. at 291. Here, shoddy or dishonest work in the laboratory could have produced a match between the DNA recovered from the shotgun and Sedillo’s DNA.
In this case, several safeguards that the plurality relied on in Williams are absent. First, the district court failed to provide careful jury instructions regarding how the jury should weigh Willard’s test results, and the district court did not consider the prejudicial effect of permitting Zais to testify regarding Willard’s underlying analysis. Second, Zais in her testimony referred to Willard’s notes and the laboratory report. In addition, there is no strong circumstantial evidence in this case regarding the reliability of Willard’s work. Applying the plurality’s first line of reasoning in Williams suggests that Zais’s testimony would violate the Confrontation Clause.
Whether the Laboratory Report Was Testimonial
Based on a second, independent reason, the plurality in Williams concluded that *690the expert witness’s statements in that case did not violate the Confrontation Clause because the Cellmark report was not testimonial. The plurality found that the Cellmark report was not testimonial because the report “plainly was not prepared for the primary purpose of accusing a targeted individual,” nor was the primary purpose “to create evidence for use at trial.” Id. at 2243. In Williams, when the sample from the victim was sent to Cellmark, the primary purpose of the DNA test was “to catch a dangerous rapist who was still at large,” as the defendant was neither in custody nor under suspicion at the time. Id. Under the facts presented, the plurality concluded that “there was no ‘prospect of fabrication’ and no incentive to produce anything other than a scientifically sound and reliable profile.” Id. at 2244 (quoting Bryant, 131 S.Ct. 1143, 1157 (2011)).
By contrast, when applying the plurality’s reasoning in Williams, the laboratory report in this case would be considered testimonial. Here, Sedillo had been identified as a suspect when a sample of his DNA, along with the shotgun, were sent to the Scientific Laboratory Division of the New Mexico Department of Public Safety. The primary purpose of the DNA test was to accuse a targeted individual and to create evidence for use at trial. Willard knew that if her test results were inculpatory, the DNA profile she produced would be used by the prosecution in its case against Sedillo. See III R. 291 (noting that Willard in this case “was informed that this case involved Paul Reyes Sedillo”). In Williams, the plurality concluded that “there is no real chance that ‘sample contamination, sample switching, mislabeling, or fraud’ could have led Cellmark to produce a DNA profile that falsely matched petitioner.” Id. at 2244. The plurality explained that it is “beyond fanciful” that “shoddy lab work would somehow produce a DNA profile that just so happened to have the precise genetic makeup of petitioner, who just so happened to be picked out of a lineup by the victim.” Id. In this case, however, it is not beyond fanciful that sample contamination, sample switching, mislabeling, or fraud could have produced test results that matched Sedillo’s DNA to the DNA sample found on the shotgun. Applying the plurality’s primary purpose test to this case, the laboratory report here would be testimonial in nature, and Zais’s testimony disclosing its contents would violate the Confrontation Clause.3
Justice Thomas in Williams disagreed with both of the plurality’s reasons. Instead, Justice Thomas concluded that the statement at issue was not testimonial because it did not bear any “indicia of solemnity.” Williams, 132 S.Ct. at 2259 (Thomas, J., concurring in judgment). Justice Thomas explained that the report in Williams was not testimonial because it “lacks the solemnity of an affidavit or deposition, for it is neither a sworn nor a certified declaration of fact,” nor does it “attest that its statements accurately reflect the DNA testing processes used or *691the results obtained.” Id. at 2260 (Thomas, J., concurring in judgment). Justice Thomas distinguished the report in Williams, which was signed by two reviewers who neither purported to have performed the DNA testing nor certified the accuracy of those who did, with the laboratory reports that were determined to be testimonial in Melendez-Diaz and Bullcoming. In Melendez-Diaz, the laboratory reports were sworn to before a notary public by the analysts who conducted the test, and in Bullcoming, the report included a “Certificate of Analyst” signed by the analyst who conducted the test. Id.
Applying Justice Thomas’s indicia of solemnity test to this case, it is unclear whether the laboratory report would be testimonial. The report was not admitted into evidence in district court and is not part of the record on appeal. Thus, it is unclear whether the report bears sufficient “indicia of solemnity” to be considered testimonial under Justice Thomas’s test.
Nonetheless, under Williams, Zais’s statements at trial disclosing the substance of the laboratory report would violate the Confrontation Clause. Zais’s statements in this case would violate the Confrontation Clause under both reasons provided by the plurality in Williams. Her' statements would also violate the Confrontation Clause according to the reasoning provided by the dissent in Williams. See id. at 2267 (arguing that the type of expert testimony in Williams “is functionally identical to the ‘surrogate testimony’ ... in Bull-coming”) (Kagan, J., dissenting). Based on the facts of this case, I would conclude that Zais’s testimony violated the Confrontation Clause.
Ill
Accordingly, I would reverse and remand with directions to vacate Sedillo’s conviction.
. At the close of the government’s case-in-chief during trial, Sedillo moved for a judgment of acquittal pursuant to Fed.R.Crim.P. 29, arguing that there was insufficient evidence to establish either actual or constructive possession. Ill R. 532-33. Sedillo renewed his motion for judgment of acquittal at the close of all evidence. Id. at 541-42. The district court denied Sedillo’s motion for judgment of acquittal. The district court found that the government had presented sufficient evidence, particularly the DNA evidence that was recovered from the shotgun, to support a theory of actual possession. Id. at 556. Relying on United States v. Hishaw, 235 F.3d 565 (10th Cir.2000), the district court also found that "the government ha[d] presented sufficient evidence to support at least the plausible inference that [Sedillo] had knowledge and access to the weapon.” Id. at 565.
. This court applied the Williams decision in United States v. Pablo, 696 F.3d 1280 (10th Cir.2012), which was decided after the prior opinion in that case, United States v. Pablo, 625 F.3d 1285 (10th Cir.2010), was vacated and remanded by the Supreme Court in light *687of Williams. See Pablo v. United States, - U.S. -, 133 S.Ct. 56, 183 L.Ed.2d 699 (2012). The defendant in Pablo was convicted of rape, kidnapping, assault, and carjacking. At trial, the government called a lab analyst to testify as an expert witness regarding DNA analysis and serology analysis performed by the state crime lab. The defendant argued that his confrontation rights were violated when the witness introduced out-of-court testimonial statements contained in the laboratory reports. Pablo, 696 F.3d at 1284. The defendant failed to raise his Confrontation Clause objection in the district court, and so the court reviewed his claim for plain error. Id. at 1287. Upon remand from the Supreme Court, the court explained that “we need not decide the precise mandates and limits of Williams, to the extent they exist.” Id. at 1289. Instead, the court emphasized that under plain error review, "we cannot say the district court plainly erred in admitting [the witness's] testimony, as it is not plain that a majority of the Supreme Court would have found reversible error with the challenged admission." Id. at 1291. Because the standard of review here is harmless beyond a reasonable doubt, as opposed to plain error review, Pablo provides little guidance for this case.
. A finding that the laboratory report in this case was testimonial would be consistent with the First Circuit’s recent interpretation of the Williams decision. In United States v. Cameron, 699 F.3d 621 (2012), the First Circuit held that certain reports generated by Yahoo! were testimonial and violated the Confrontation Clause. In distinguishing Cameron from Williams, the First Circuit noted that “[n]obody at Yahoo! who was involved in creating the [reports] could possibly have believed that the [reports] could be other than 'incriminating.' " Id. at 647. The First Circuit held that the reports at issue were testimonial because “Yahoo! created these Reports after its own employees had already concluded that a crime had been committed,” and that "Yahoo! then sent these Reports to an organization that forwards such reports to law enforcement.” Id.