State v. Leuluaialii

Full title: THE STATE OF WASHINGTON, Respondent, v. KENNETH JOHN LEULUAIALII, ET AL.…

Court: The Court of Appeals of Washington, Division One

Date published: Oct 13, 2003

118 Wn. App. 780 (Wash. Ct. App. 2003)

118 Wash. App. 780

77 P.3d 1192

Opinion

Nos. 43507-8-I; 43769-1-I.

October 13, 2003.

David B. Koch (of Nielsen, Broman Koch, P.L.L.C.) and Sharon J. Blackford (of Washington Appellate Project), for appellants.

Norm Maleng, Prosecuting Attorney, and Brian M. McDonald and Catherine M. McDowall, Deputies, for respondent.

GROSSE, J.

The methodology and technique for extracting and identifying deoxyribonucleic acid (DNA) in plants and animals is relatively well established, as is the methodology and technique for determining the probability and extent of a relationship between two samples of human DNA, including matching the DNA of a sample to a specific human being. However, the study of canine DNA has not progressed to the point of the study of human DNA sufficient to permit an expert to testify to a match between a sample and a specific dog. It was error for the trial court to have admitted such testimony, but an error that was harmless in the context of this case. Likewise, the erroneous instruction on accomplice liability was harmless.

The convictions of the appellants are affirmed and the case remanded for resentencing.

FACTS

In the early morning hours of December 9, 1996, Jay Johnson (Johnson) and Raquel Rivera (Rivera) were shot to death in their home. The victims' dog Chief received gunshot wounds and later died from complications related to the wounds. Johnson's and Rivera's assailants were searching for drugs and money and were members of a gang that included George Leuluaialii (Leuluaialii) and Kenneth Tuilefano (Tuilefano). Leuluaialii was charged and convicted of two counts of aggravated murder in the first degree, and one count of animal cruelty in the first degree. Tuilefano was charged and convicted of two counts of murder in the first degree.¹

1 Although Tuilefano argues otherwise, the State charged him with four alternate means of first degree murder several months prior to trial: (1) premeditated murder, (2) felony murder with first degree robbery as the predicate felony, (3) felony murder with second degree robbery as the predicate felony, and (4) felony murder with residential burglary as the predicate felony. Thus, he received adequate notice of all charges.

Other individuals present claimed that although they participated in the search of the victims' home for drugs and money, they did not participate in the murders of Johnson and Rivera or in the shooting of the dog. These individuals were Tuiafea Tauiliili Jr. (Tauiliili), Charles Niko (Niko), Malini Fosi (Fosi), and Sheila Harris (Harris), Fosi's fiancé. Tauiliili, Niko, and Fosi agreed to give testimony to the State in return for reduced charges. Additional independent eyewitnesses from the neighborhood also testified at trial.

The witnesses who observed the events testified that around 8:30 or 8:45 A.M. on the day in question, Leuluaialii, wearing a dark coat, something over his head, and carrying a gun, kicked in the victims' door and began shooting. Leuluaialii first shot the victims' dog Chief. He then entered the residence and more shots were heard being fired inside. There were also sounds of a woman screaming. Tuilefano, also wearing something over his head, followed Leuluaialii inside before any of the other individuals. Tauiliili and Niko claimed that they entered the apartment only after this initial activity. Harris and Fosi testified that they remained outside. Witnesses who did not observe all of the events saw Leuluaialii near the victims' home around the time of the murders and heard gunshots.

No one witnessed the shooting of Rivera. Tauiliili testified that when he entered the bedroom, Rivera had already been shot, and Johnson had been shot in the stomach. Tauiliili stated he saw Leuluaialii holding a gun, kick and hit Johnson with the gun, and then saw Leuluaialii hand the gun to Tuilefano, telling him to "smoke this motherfucker." Niko testified that he saw Tuilefano shoot Johnson twice in the thigh.

Ballistic evidence revealed that Rivera was shot three times: once in the chin, once in the neck, and once in the hand. Autopsy results revealed that Johnson suffered two gunshot wounds to his abdomen and two to his legs. Although the murder weapon was not recovered, the State presented evidence that the bullets recovered from the victims' bodies were fired from the same gun used days earlier in an assault on an individual named Johnny Paik. Paik identified Leuluaialii in a police photo montage and in court as the person who shot at him. Fosi testified that he was with Leuluaialii during the assault, handed Leuluaialii the gun, and saw him shoot at someone matching Paik's description.

Niko and Tauiliili testified that Leuluaialii, Tuilefano, Niko, and Tauiliili searched the home for drugs and money. Tuilefano and Leuluaialii left the scene in a red Camaro which various witnesses saw parked near the victims' home around the time of the murders. After the killings, Leuluaialii, Tuilefano, Niko, and Tauiliili headed to the Biltmore Hotel in Tacoma. Tuilefano and Niko testified that they soon returned to the scene in the Camaro and were arrested. Leuluaialii was arrested the next morning at the hotel.

Both Tuilefano and Leuluaialii claimed that they were not present during the crimes. Although he did not testify in his defense, Tuilefano maintained that he never entered the victims' home and he did not participate in any of the crimes. Leuluaialii testified and denied being anywhere near the victims' home on the day of the murders. Leuluaialii's mother and father testified that on the morning of the murders he was asleep at home on the couch when they left for work, just before 5 A.M. and 8 A.M., respectively. His mother testified that when she called home between 8:45 A.M. and 9 A.M., Leuluaialii answered the phone and he sounded like he was still asleep. Leuluaialii claimed that he did not wake until 1 P.M. He then went to the Biltmore Hotel where he drank and passed out. He testified that he did not awaken until he was arrested.

Three black jackets were recovered at the Biltmore Hotel where Leuluaialii was arrested.² To counter the appellants' claims that they were not present during the murder, the State provided the following physical evidence at trial:

2 The State claimed that one of the jackets found at the hotel belonged to Leuluaialii and another jacket belonged to Tuilefano. Leuluaialii denied that any of the jackets found in the hotel room belonged to him.

• Leuluaialii's left shoe (SPD # 27) had characteristics that would have left the print on the victims' front door and could not be eliminated as a source of the print.

• Several hairs found on Leuluaialii's and Tuilefano's shoes (SPD # 27 and SPD # 34) had characteristics similar to Chief's hair.

• A single hair found on a jacket attributed to Leuluaialii (SPD # 50) also had characteristics similar to Chief's hair.

• A piece of cloth similar to the hood or mask a witness testified that Leuluaialii wore was recovered at the scene.

• Human blood was found on Tuilefano's pants (SPD # 32) and on the jacket attributed to Tuilefano (SPD # 49).

• Both human blood and dog blood were found on the jacket attributed to Leuluaialii (SPD # 50).

The State also conducted DNA testing of the canine and human DNA evidence on several items of clothing with the assistance of three laboratories: PE Zoogen, a division of PE Applied Biosystems; Forensic Science Associates; and Cellmark Diagnostics (Cellmark). Prior to trial, Tuilefano filed a motion requesting a Frye ³ hearing and an ER 702 hearing for the DNA testing done by Cellmark and PE Zoogen. Leuluaialii joined in Tuilefano's motion.

3 Frye v. United States, 293 F. 1013, 1014 (D.C. Cir. 1923) (to be admissible, novel scientific evidence must generally be accepted by the relevant scientific community).

With regard to the canine DNA tested by PE Zoogen, the appellants primarily argued that no standardized testing was available for canine DNA and that a Frye hearing was necessary with regard to PE Zoogen's test. The appellants also argued that the polymerase chain reaction (PCR) DNA testing used by Cellmark utilized various loci not yet accepted in the scientific community, and the lab's procedures were unreliable and inadmissible under ER 702. Tuilefano focused on the DQ alpha loci used by Cellmark, but also included the LDLR, GYPA, HBGG, D7S8, and GC loci in his motion for a Frye hearing.

PE Zoogen tested three articles of clothing for the State which contained canine blood: two of three black jackets found in the Biltmore Hotel room attributed by the State to Tuilefano and Leuluaialii, and Leuluaialii's pants. PE Zoogen used Chief's sample blood and 10 DNA markers in its testing. PE Zoogen's founder, Dr. Joy Halvorson, concluded that for all items tested at least 9 of the 10 markers matched the DNA found in various stains on the items. The two jackets attributed to Tuilefano and Leuluaialii had some stains that matched all 10 markers. Dr. Halvorson testified that with a 9-marker match, the probability of finding another dog with Chief's DNA profile was 1 in 18 billion; with a 10-marker match, the probability was 1 in 3 trillion. PE Zoogen used Chief's sample blood and 10 DNA markers, or loci, in its testing. Seven of the markers were developed by PE Zoogen in its mixed breed database which consisted of approximately 100 dogs, most of which came from a local dog shelter in Los Angeles. The other three were developed by the Fred Hutchinson Cancer Research Center in Seattle, Washington.

A chart, "Summary of DNA Evidence," was included in the State's brief and we reproduce it here:

LABORATORY ITEM DESCRIPTION RESULTS -------------------------------------------------------------------------- Forensic Science Assocs. SPD 49 Tuilefano's mixed stains; unable jacket to determine contributors

Forensic Science Assocs. SPD 50 Leuluaialii's mixed stains; unable jacket to determine contributor sexcept to exclude Leuluaialii as to area F

Cellmark SPD 32 Tuilefano's mixed stain; Raquel pants Rivera potential primary contributor (1 in 28,000); Jay Johnson, Tuilefano and Leuluaialii potential minor contributors

Cellmark SPD 49 Tuilefano's Niko and Fosi jacket excluded; Raquel Rivera, Jay Johnson and Tuilefano possible contributors

Cellmark SPD 50 Leuluaialii's non-human DNA jacket indicated

PE Zoogen SPD 31 Leuluaialii's matches Chief jeans

PE Zoogen SPD 49 Tuilefano's matches Chief jacket

PE Zoogen SPD 50 Leuluaialii's matches Chief jacket

Tuilefano and Leuluaialii contested the results of both the canine and human DNA tests at trial and presented their own expert witnesses. Prior to closing argument, Tuilefano moved to strike all of the State's DNA evidence. Leuluaialii did not join in the motion. Tuilefano argued that various test results were inadmissible because the testing methods were not scientifically accepted, samples were mislabeled, various results were unreliable, and statistical probabilities were not given with some test results.

The court admitted Cellmark's test results, concluding that its PCR testing, and use of polymarkers, short tandem repeats (STRs), and the DQ alpha loci, were also generally accepted in the scientific community. The trial court ruled that the methodologies used in PE Zoogen's analysis of the canine DNA, utilizing PCR DNA testing of polymarkers on STRs sections of the canine DNA strand, was a method of DNA testing generally accepted in the scientific community. It concluded that a Frye hearing was not necessary. The court also found the appellants failed to make a prima facia showing that an ER 702 hearing was necessary for either the human or canine DNA evidence.

DISCUSSION

1. DNA Evidence.

There is no dispute in this case with regard to the Frye standard or that it requires both accepted theory and a valid technique to implement that theory.⁴ There is also no serious dispute in this case but that the testing with regard to human DNA was done by means of accepted theory. The parties' positions diverge with regard to technique. However, as to technique, we believe the arguments about human DNA are better focused by reference to ER 702 and its two-part test: (1) does the witness qualify as an expert and (2) will the testimony be helpful to the jury.⁵ Here, the appellants argue that the testing procedures for human DNA were so flawed as to be unreliable and thus should have been excluded as not helpful.⁶ However, a dispute over the validity of particular procedures generally goes to the weight of the evidence, not its admissibility.⁷ In short, the issue for the trial court was one of degree and focused on the showing made by the defense to support its request for an ER 702 hearing. On the record, we agree with the trial court that the dispute over human DNA evidence goes to weight not admissibility. While an ER 702 hearing would have been preferable, the defense showing on this point was scant.

4 State v. Riker, 123 Wn.2d 351, 359, 869 P.2d 43 (1994).

5 See State v. Cauthron, 120 Wn.2d 879, 890, 846 P.2d 502 (1993).

6 See State v. Russell, 125 Wn.2d 24, 50-51, 882 P.2d 747 (1994); State v. Copeland, 130 Wn.2d 244, 270-71, 922 P.2d 1304 (1996).

7 See Copeland, 130 Wn.2d at 270-71 (DNA evidence properly admitted despite defense challenges of laboratory error, lack of proficiency testing, and inadequacy of database).

But, with respect to the canine DNA evidence, we must differ with the trial court. The trial court ruled that the methodologies used in PCR testing, including the PCR testing used by PE Zoogen which analyzed STRs, were generally accepted in the scientific community and that no Frye hearing was needed. The trial court allowed the State's witness to testify that for all articles of clothing tested for canine blood, at least 9 of 10 loci matched Chief's DNA, and 10 loci matched on one article of clothing. The court then admitted the State's testimony that the probability of having the DNA come from another dog other than Chief was "1 in 3 trillion" for a 10 loci match, and "1 in 18 billion" for a 9 loci match.

The evidence here clearly involved novel scientific theory: the forensic identification with high statistical probabilities of a specific dog through analysis of canine DNA. There are no published United States cases that involve the use of canine DNA markers for forensic purposes or examine the validity of the specific markers used here. A Frye hearing was absolutely necessary in the present case.

This court reviews the trial court's decision to admit or exclude scientific evidence under Frye de novo.⁸ The reviewing court may consider other evidence not in the record, including scientific and law review articles and decisions from other jurisdictions, to determine whether (1) the scientific theory has general acceptance in the scientific community, (2) the techniques and experiments that currently exist can produce reliable results and are generally accepted by the scientific community, and (3) the laboratory performed the accepted scientific techniques in this case.⁹

8 Cauthron, 120 Wn.2d at 887.

9 Russell, 125 Wn.2d at 41; Cauthron, 120 Wn.2d at 887-89.

We have examined the record, existing case law on DNA evidence, scientific literature on the canine genome, and the documents submitted at our request by the parties. Based on this information, we are not convinced that forensic canine DNA identification is a theory that has received general acceptance in the scientific community, or that reliable techniques or experiments exist to identify individual canines for forensic purposes.

In State v. Russell, ¹⁰ the defendant challenged the general acceptance of the PCR technique using a DQ alpha genetic marker system. The trial court heard testimony from experts that over 30 private and government laboratories were either in the process of implementing DQ alpha typing or already using it (including the Federal Bureau of Investigation (FBI)).

10 State v. Russell, 125 Wn.2d 24, 882 P.2d 747 (1994).

Following this testimony, the trial judge found the results of PCR testing at the HLA [human leukocyte antigen system] DQ alpha locus admissible under the Frye standard. The judge found that the underlying principle and techniques of PCR had been generally accepted by the scientific community, and added that DQ alpha testing and typing had gone sufficiently beyond the experimental stage to gain general acceptance in the scientific community. She observed further that the DQ alpha gene has been subjected to considerable scientific study, especially in the fields of immunology and medicine. The variations of the gene are well known, readily identified, and easily distinguished, making this gene an appropriate genetic marker for forensic use. She also noted that the population frequencies of the various genotypes occurring at the DQ alpha locus were not contested. The judge concluded that the fact that a scientific procedure might yield a false result if not performed properly did not render it inadmissible and that any problems associated with PCR testing at the DQ alpha locus went to the weight to be given the evidence.¹¹

11 Russell, 125 Wn.2d at 44.

Russell challenged these conclusions. The Supreme Court affirmed the trial court holding that the PCR amplification system for testing DNA at the HLA DQ alpha locus was generally accepted in the relevant scientific community. The PE Zoogen markers are in marked contrast.

In order to satisfy its concerns over the product rule as a means to estimate genetic profile frequencies, the Supreme Court looked to the extensive studies conducted worldwide by the FBI to demonstrate that human databases are in Hardy-Weinberg and linkage equilibrium.¹² We do not have such information for canines.

12 Hardy-Weinberg equilibrium ensures "that the statistical calculations are based on a truly random population — one which mates randomly and thus mixes the gene pool evenly." Cauthron, 120 Wn.2d at 902. Linkage equilibrium ensures "that each of the probes used detects an allele which is independent of the other alleles tested. That is, in calculating the statistics, the scientist attempts to ensure that the various sites tested are not related to each other." Cauthron, 120 Wn.2d at 902. Compare, Cauthron, 120 Wn.2d at 905-06 with Copeland, 130 Wn.2d at 266.

Although the general underlying principles of PCR testing of human DNA are accepted by the scientific community, all of the Washington cases that address the admissibility of DNA evidence recognize that human DNA and loci have been subjected to substantial testing and validation by the scientific community.¹³ Only a small percentage of the overall number of sites on the human DNA are variable, 0.1 to 0.3 percent of 3 billion, and scientific testing has resulted in the identification of highly variable, polymorphic portions of the DNA called loci or markers.¹⁴

13 See, e.g., Copeland, 130 Wn.2d at 261-70; Russell, 125 Wn.2d at 37-55; Cauthron, 120 Wn.2d at 891-96. See also NAT'L RESEARCH COUNCIL, COMM. ON DNA FORENSIC SCI., THE EVALUATION OF FORENSIC DNA EVIDENCE 127 (1996) (discussing the international nature of the human DNA database compiled by the FBI as of 1993).

14 Cauthron, 120 Wn.2d at 900-01.

These loci contain alleles, or alternate forms of a gene at the specific loci.¹⁵ Scientific testing has also determined the relative probability estimates of these polymorphic markers and alleles in humans. Thus, the recurrence of several highly polymorphic markers in a tested human can be estimated with reasonable probabilities under the product rule.¹⁶

15 LORNE T. KIRBY, DNA FINGERPRINTING: AN INTRODUCTION (1990).

16 To calculate probabilities based on the product rule, a scientist first collects the data pertinent to each human allele or DNA marker being used as a comparison and creates a database. Based on the statistics of occurrence of these individual DNA markers in the database, probabilities of the occurrence of several alleles in a particular test human can be calculated.

For instance, allele A may be found in 1 of every 10 people; allele B found in 1 of 20; and allele C found in 1 of 5. Under the product rule, if there is a match for each allele, the expert can multiply (1/10 × 1/20 × 1/5) to achieve the result that only 1 person in 1,000 will match all three sites.

Cauthron, 120 Wn.2d at 901. Calculation of probability estimates under the product rule rests on the assumption that the loci, or alleles, are in both Hardy-Weinberg and linkage equilibrium.

To determine whether identifying a dog through canine DNA with the same forensic accuracy of identifying a human is accepted in theory and practice by the relevant scientific community and thus admissible here, the trial court should have established: (1) what was known about the nature and extent of genetic variation in canines, (2) whether the scientific community generally accepted that the specific loci used by PE Zoogen in the present case were highly polymorphic and appropriate for forensic use, and (3) whether the scientific community generally accepted that PE Zoogen's calculations of a random match were based on accurate probability estimates.¹⁷

17 See, e.g., Frye, 293 F. at 1014 ("while courts will go a long way in admitting expert testimony deduced from a well-recognized scientific principle or discovery, the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs"); Russell, 125 Wn.2d at 37-40, 42-51, 54 (recognizing that DNA identification of humans is generally accepted by the scientific community and is based on knowledge about the extent of genetic variation in humans, identification of polymorphic DNA loci, and accurate probability estimates, all developed through years of extensive research and validation by the scientific community); Cauthron, 120 Wn.2d at 895-98 (there is nothing controversial about the theory underlying human DNA typing, or the practice of restricted fragment length polymorphism typing of human DNA, because the principles are well-established). See also George Sensabaugh D.H. Kaye, Non-Human DNA Evidence, 39 JURIMETRICS J. 1 (Fall 1998).

We have determined through our own examination of the record and the publications available in the scientific community that canines have genetic diversity similar to that observed in humans.¹⁸ Further, portions of the canine DNA are highly polymorphic and thus would yield polymorphic loci and alleles that might be appropriate for forensic use.¹⁹

18 See, e.g., Mark W. Neff et al., A Second-Generation Genetic Linkage Map of the Domestic Dog, Canis familiaris, 151 GENETICS 803 (Feb. 1999); C. Vila, J.E. Mondato R.K. Wayne, Phylogenetic Relationships, Evolution, and Genetic Diversity of the Domestic Dog, 90(1) J. OF HEREDITY 71 (1999); L.V. Francisco et al., A class of highly polymorphic tetranucleotide repeats for canine genetic mapping, 7 (5) MAMMALIAN GENOME 359-62 (1996).

19 See, e.g., S. Muller et al., Use of Canine Microsatellite Polymorphisms in Forensic Examinations, 90(1) J. OF HEREDITY 55-56 (1999); Neff, supra note 18; M. Olivier, M.A. Meehl G. Lust, Random Amplified Polymorphic DNA (RAPD) Sequences as Markers for Canine Genetic Studies, 90(1) J. OF HEREDITY 78, 81 (1990); P.M. Schneider, Y. Seo C. Rittner, Forensic mtDNA hair analysis excludes a dog from having caused a traffic accident, 112(5) INT'L J. LEGAL MED. 315-16 (1999).

However, we were not able to conclusively determine whether all of the loci used in this case were polymorphic markers. More importantly, we also were not able to determine whether PE Zoogen's frequency estimates of its loci and alleles were accurate. The defense presented several expert witnesses who vehemently contested the validity of Zoogen's canine DNA markers.²⁰ PE Zoogen indicated at trial that it would soon publish studies of its loci and their frequency estimates, but we have found no such studies and the State provided none despite a request by this court.

20 Dr. Lawrence Mueller, a professor at the University of California-Irvine who has taught population statistics, testified that all of the markers PE Zoogen used failed to conform to Hardy-Weinberg equilibrium, and the errors were "pretty severe." He testified that more than 30 percent of the various loci used by PE Zoogen as markers showed dependence, while one would typically expect to see no more than five percent dependence for proper linkage equilibrium. Dr. Mueller concluded that many of the dogs used in the all-breed database were related and PE Zoogen should not have used the product rule to establish statistical probabilities of DNA matches to other canines. Dr. William Shields, a professor of biology at the State University of New York College, testified that while there has been extensive research on forensic human DNA typing, there has been less study or understanding of forensic animal typing. He asserted that the necessary validation studies for animal DNA typing have not been done and that PE Zoogen's markers had not been "characterized," or subjected to a series of tests required for their acceptance as reliable forensic DNA markers. Further, because canines are inbred, Dr. Shields stated that the frequency of common alleles is greater and probability estimates might be undermined. Dr. Libby, a molecular geneticist, echoed these concerns and concluded that PE Zoogen's markers were not ready for forensic application.

We recognize the State's contention that probability estimates in the trillions for human DNA markers are now accepted in court.²¹ Indeed, probability estimates of occurrences of various human DNA profiles may exceed the human population because of the extent of knowledge about human DNA and the frequency of certain polymorphic alleles.²² However, not as much is known about polymorphic loci in canine DNA and the probability of their occurrence as is known about human polymorphic loci. Current canine DNA testing and mapping focuses on the goals of paternity testing, pure breed testing, and cancer and disease research studies.²³ There is little indication that polymorphic loci and alleles in canine DNA have been sufficiently studied such that probability estimates are appropriate for the forensic use applied in this case.

21 See, e.g., State v. Buckner, 133 Wn.2d 63, 66, 941 P.2d 667 (1997).

22 NATIONAL RESEARCH COUNCIL, supra note 13, at 33.

23 See, e.g., Neff, supra note 18.

The lack of published studies on PE Zoogen's canine DNA markers and their probability estimates, and the disagreement of the experts at trial about their accuracy, leaves this court disinclined to accept the statistics given at trial. Further, we note the existence of a case from British Columbia that connected a defendant to a crime scene through the use of PE Zoogen's markers.²⁴ PE Zoogen claimed, as here, that the markers used were highly polymorphic and estimated a 1-in-3 billion chance existed that the canine DNA tested on the defendant's pants came from any other dog than the dog killed at the crime scene.

24 Gary G. Shutler et al., Removal of a PCR Inhibitor and Resolution of DNA STR Types in Mixed Human-Canine Stains from a Five Year Old Case, 44(3) J. FORENSIC SCI. 623 (1999).

Nine of the ten loci identified by PE Zoogen in the DNA of the dog in British Columbia were also found in Chief's DNA in the present case. Six of these loci were developed by PE Zoogen. Chief was a mixed breed dog with the predominant breed being pit bull or one of the bull terrier breeds,²⁵ and the dog in the British Columbia case was a terrier, yet they had significant genetic similarities in terms of these nine markers. The British Columbia case referenced in the article makes us question both the polymorphic nature of PE Zoogen's markers and the statistics used to estimate their frequency of occurrence in the canine population.²⁶

25 The veterinarian who treated Chief for his injuries testified as to his breed as follows:

I think the predominant breed was probably pit bull or one of the bull terrier breeds. He certainly is crossed with another large breed dog. It could have been Labrador or German Shepard or even Great Dane. It's a little hard to tell. But he was taller than a standard pit bull.

Report of Proceedings (July 28, 1998) at 23.

26 We also located two other reports which examined six of the PE Zoogen loci used in the present case: PEZ1, PEZ3, PEZ5, PEZ6, PEZ8, and PEZ12. One case report recognized that exact forensic application of the loci used required population studies of the local canine populations and data of allele frequencies of the loci examined. The other asserts that probability estimates for the loci examined had been published, but gives no cites or sources for published studies of allele frequencies for the loci. See, e.g., Z. Padar et al., Canine STR analyses in forensic practice: Observation of a possible mutation in a dog hair, 116(5) INT'L J. LEGAL MED. 286-88 (Oct. 2002) (canine DNA loci used to identify two possible perpetrators of a canine attack); Z. Padar et al., Canine microsatellite polymorphisms as the resolution of an illegal animal death case in a Hungarian Zoological Gardens, 115(2) INT'L J. LEGAL MED. 79-81 (2001) (canine DNA loci used to eliminate the zoo's German shepherd guard dogs as suspects in an animal death case).

We are aware that canine DNA STR loci, like human STRs, may contain many separate identifiable alleles. Thus, identification of the same nine loci on two canine DNA samples may not be statistically improbable if the identified alleles within the loci differ.²⁷ In fact, at least six of the PE Zoogen loci used here have since been characterized as multi-allele sites.²⁸ Further, at least three of the loci used here are characterized as unlinked.²⁹ This supports the possible conclusion that these three particular loci may be good candidates for probability estimates and studies.³⁰ However, the lack of any studies to substantiate the statistical probabilities of PE Zoogen's loci and alleles used here, and thus their accuracy in identifying a specific dog with a "1 in 18 billion" or "1 in 3 trillion" probability, leaves us with little choice but to conclude that a Frye hearing was necessary.

27 Compare NATIONAL RESEARCH COUNCIL, supra note 13, at 33-34, 70, 118-19; Muller, supra note 19, at 56.

28 Neff, supra note 18, at 808 (PEZ1, PEZ3, PEZ5, PEZ6, PEZ8, PEZ12).

29 Neff, supra note 18, at 814-16 (PEZ3, PEZ5, and PEZ6, on linkage groups L23, L9 and L11, respectively).

30 See, e.g., NATIONAL RESEARCH COUNCIL, supra note 13, at 25-35 (linkage equilibrium for purposes of probability estimates under the product rule relies, in part, upon nonlinked alleles that have less tendency to be inherited together); LORNE T. KIRBY, DNA FINGERPRINTING: AN INTRODUCTION XV (1990).

In sum, the studies presently available on canine polymorphic loci, or the frequency of their occurrence in the canine population, are insufficient for specific markers to be accurate forensic indicators as they were used in the present case. Because PE Zoogen has not yet published sufficient data to show that its DNA markers and associated probability estimates are reliable, we would suggest that other courts tread lightly in these waters and closely examine canine DNA results before accepting them at trial.³¹

31 This court notes that PE Zoogen and its related corporate identities, PE AgGen, Celara AgGen, and Perkin-Elmer Corporation, appear to be the only entities offering canine DNA testing for forensic purposes in criminal trials.

2. DNA Harmless Error.

The error in admitting the canine DNA evidence was harmless. It was not an error of constitutional magnitude and we are convinced that the error did not affect the outcome of the trial within reasonable probabilities.

We will not reiterate the evidence we recite at the beginning of this opinion. The defense was that neither Leuluaialii nor Tuilefano was present, but absent Leuluaialii's testimony and that of his parents, there is overwhelming evidence that he was present and that he was the principal actor. As for Tuilefano, if he was there, there is no reasonable doubt but that he was at least an accomplice if not a participant. Thus, we turn to the debate over the charges and instructions as to Tuilefano.

The remainder of this opinion has no precedential value. Therefore, it will be filed for public record in accordance with the rules governing unpublished opinions.

KENNEDY and SCHINDLER, JJ., concur.