Summary
holding that denial of access to the FBI lab's DNA data bases did not make its test results inadmissible
Summary of this case from State v. UnderdahlOpinion
No. C9-92-2562.
July 27, 1993.
Appeal from the District Court, Olmsted County, Gerard Ring, J.
Hubert H. Humphrey, III, Atty. Gen., St. Paul, Raymond F. Schmitz, Olmsted County Atty., Rochester, Stephen Redding, Asst. Hennepin County Atty., Minneapolis, for appellant.
Charles Nicholas, Rochester, William Kennedy, Hennepin County Public Defender, Patrick J. Sullivan, Asst. Hennepin County Public Defender, Minneapolis, for respondent.
Considered and decided by ANDERSON, C.J., HUSPENI and DAVIES, JJ.
OPINION
The prosecution appeals a pretrial order excluding statistical evidence and nonstatistical opinion testimony on the significance of a DNA "match." Respondent has filed a cross-appeal challenging the admission of the DNA test results. We affirm in part, reverse in part, and remand.
FACTS
Respondent Daryl Duane Alt is charged with first degree criminal sexual conduct, first degree assault, and attempted murder for the beating, rape, and attempted strangulation of a 15-year-old female. The alleged assault by Alt occurred in August 1990. In September 1990, biological samples from the victim's clothing worn during the attack, a vaginal swab taken during the victim's rape protocol examination, and body fluids from Alt and the victim were sent to the Federal Bureau of Investigation (FBI) laboratory in Washington, D.C. for deoxyribonucleic acid (DNA) testing. Using standard DNA Restriction Fragment Length Polymorphism (RFLP) analysis, the FBI compared the DNA profile from the crime scene biological samples against the body fluid samples taken from Alt and the victim.
The FBI analysis demonstrated that Alt's known DNA profile matched semen stains taken from the victim's clothing and vaginal swab across the four loci — sample fragments of DNA — tested. The FBI also determined the statistical frequency with which one could expect to find such a match among samples taken from three of its four major data bases.
An FBI analyst testified that the FBI maintains four main data bases (Caucasian, Hispanic, African American, and American Indian) developed from pools of DNA samples taken from different people.
The state received the FBI DNA test results in April 1991. At that time, the court proceedings had been stayed pending the state's appeal of the district court's order suppressing evidence. This court affirmed the district court's order in May 1991. See State v. Alt, 469 N.W.2d 732 (Minn.App. 1991).
After the appeal, the state sought to admit the results of the DNA analysis and the FBI probability figures showing the statistical significance of the match. The state agreed to submit the probability evidence of a match at each locus separately, relying on State v. Joon Kyu Kim, 398 N.W.2d 544 (Minn. 1987). The state agreed not to introduce the composite statistical frequency evidence on all four loci, as permitted under Minn.Stat. § 634.26 (1990).
While we do not speculate on the state's reasons for agreeing not to seek admission of the evidence under this statute, we note that the supreme court has recently called into question the legislature's authority to create an exception for DNA evidence to the limitations of Joon Kyu Kim on the use of statistical probability evidence in criminal trials. See State v. Nielsen, 467 N.W.2d 615, 620 (Minn. 1991) (noting the discussion in State v. Willis, 332 N.W.2d 180, 184 (Minn. 1983) on the separation of powers doctrine and the power to establish rules of evidence); State v. Schwartz, 447 N.W.2d 422, 429 n. 6 (citing Willis, 332 N.W.2d at 184). We also note the legislature has recently amended Minn.Stat. § 480A.0591, subd. 6. 1993 Minn. Laws ch. 326, art. 7, § 12. However, because the state did not seek admission under section 634.26 and in light of Nielsen, we do not address the amendment's application to this case.
Alt objected and moved to have the DNA test results excluded for failure to meet the admissibility criteria for novel scientific evidence under the modified Frye standard enunciated in State v. Schwartz, 447 N.W.2d 422, 424 (Minn. 1989) (discussing Frye v. United States, 293 F. 1013 (D.C. Cir. 1923)). Alt claimed that (1) the FBI laboratory did not comply with the appropriate standards and controls, and (2) experts in the relevant scientific community would not find the limited statistical evidence based on the FBI data bases reliable and trustworthy.
A pretrial Frye hearing on the admissibility of the DNA evidence took place on nine separate days over a three-month span, ending in March 1992. Seven expert witnesses testified at the hearing, four for the state and three for Alt.
In April 1992, the National Research Council issued its report, entitled "DNA Technology in Forensic Science" (NRC Report), which was prepared by the Committee on DNA Technology in Forensic Science. The NRC Report supplements the previously issued Technical Working Group for DNA Analysis Methods (TWGDAM) guidelines for DNA testing procedures. The parties agreed that the district court could consider the NRC Report as evidence in the case.
In Schwartz, the supreme court held that the TWGDAM standards were an appropriate guide for DNA testing. See Schwartz, 447 N.W.2d at 427-28.
In May 1992, the district court issued its order limiting the use of DNA test results. The court found that FBI testing met the standards for both Frye and Schwartz and held that the DNA test results were admissible at trial. However, experts would be limited to testimony that indicated the test results did not exclude Alt as a possible source of the body fluids tested. It also determined that FBI statistical calculations did not have the general acceptance in the scientific community necessary for admissibility. Therefore, it concluded testimony regarding the probability of a match was inadmissible. In addition, the court ruled that the FBI's failure to meet the proficiency testing requirements of TWGDAM and the NRC Report should be submitted to the jury.
The state filed a motion to reconsider in June 1992. It claimed the court's findings regarding FBI compliance with TWGDAM were erroneous and the court's finding regarding the FBI statistical calculations was inconsistent with the NRC Report.
Before the hearing on the state's motion, the Minnesota Supreme Court filed its opinion in State v. Jobe, 486 N.W.2d 407, 419-20 (Minn. 1992), in which it held that the FBI laboratory's protocol established an appropriate set of standards and guidelines, despite the lack of an independent laboratory audit of FBI testing. Consequently, the state submitted a revised motion to reconsider, requesting a finding that the FBI testing procedures met the requirements of the TWGDAM guidelines and the NRC Report.
The Minnesota Supreme Court also took judicial notice of the NRC Report that had been issued after the Jobe case was tried. See Jobe, 486 N.W.2d at 420 n. 3.
The district court heard the revised motion to reconsider in August 1992, at which time the state again offered to present statistical evidence of a DNA match consistent with Joon Kyu Kim and as presented in Jobe . In December 1992, the district court denied all motions relating to the admissibility of the DNA test results and confirmed its May 1992 order.
The presentation of statistical evidence was not challenged on appeal in Jobe and, consequently, was not addressed by the supreme court.
The discussions in the district court are ambiguous with respect to what individual frequencies the state sought to introduce — individual band frequencies (seven in this case), or locus frequencies (here, four). The terms were used interchangeably in the record. The state submitted a DNA profile showing both band frequencies and locus frequencies. We construe the state's offer as one of locus frequencies, which combine the band frequencies and are the strongest probability evidence admissible under Joon Kyu Kim and Johnson.
On appeal, the state challenges the district court's failure to admit statistical frequency evidence in a limited fashion, as provided under Joon Kyu Kim, and the exclusion of nonstatistical opinion testimony about the significance of a match. Alt cross-appeals, claiming the district court's finding that the FBI's protocol and procedures were sufficient for the admissibility of the DNA test results is clearly erroneous.
After briefing in this appeal, the United States Supreme Court issued an opinion holding the Frye standard was superseded in the federal courts by the Federal Rules of Evidence. Daubert v. Merrell Dow Pharmaceuticals, Inc., ___ U.S. ___, ___ _ ___, 113 S.Ct. 2786, 2792-96, 125 L.Ed.2d 469 (1993). The state submitted a letter "brief" following the release of Daubert, and respondent has moved to strike this letter brief.
ISSUES
I. Has the state shown the district court's ruling will have a critical impact on the prosecution?
II. Did the district court properly admit the DNA test results?
III. Did the district court clearly err in ruling the probability of a random DNA match on individual loci did not meet the Frye test?
IV. Did the district court clearly err in excluding nonstatistical opinion testimony on the significance of a DNA match?
ANALYSIS DNA Testing and the RFLP Process
DNA testing for forensic purposes has great potential for use by the criminal justice system. It has the capacity to identify perpetrators of crime with a power of differentiation far beyond that of conventional serology. First used for forensic purposes in 1985, it is an outgrowth of medical diagnostic use.
The probative value of DNA testing does, however, have potential limitations. The possibilities for its misuse or abuse have raised important questions about its reliability, validity, and confidentiality. NRC Report at S-1 (prepublication manuscript). Lack of standardization, laboratory accreditation, and properly measured laboratory error rates, all noted in the NRC Report, are indices that this technology is still developing and highlight its limitations. As cases involving other scientific tests have demonstrated, even the more established forensic science techniques are not infallible. See, e.g., State v. Caldwell, 322 N.W.2d 574, 581-82, 586 (Minn. 1982) (reversing homicide conviction in part because fingerprint expert's testimony was in error). The immediate challenge facing the criminal justice system is to apply this developing scientific technology to the quest for truth in the courtroom.
As the Supreme Court noted in Daubert, "there are important differences between the quest for truth in the courtroom and the quest for truth in the laboratory." Daubert, ___ U.S. at ___, 113 S.Ct. at 2798. The scientific quest for truth involves consideration and proof of all hypotheses — both the correct and incorrect. The legal quest for truth is limited to resolving issues finally and quickly, based on testing that is less forgiving of error and which requires more extensive controls. These distinctions must be considered when determining the admissibility of DNA test results and the accompanying statistical evidence and nonstatistical opinion testimony.
Comparing the forensic use of DNA testing in the legal setting to its medical diagnostic use reveals the importance of ascertaining the forensic limits to the reliability of DNA testing procedures. As the NRC Report explains, DNA diagnostics usually involve an unlimited supply of clean tissue and/or fresh blood samples from known sources. NRC Report at S-5. This provides the analyst with the opportunity to do repeat testing to resolve unclear results and it provides quality controls or consistency checks. Moreover, population statistics are irrelevant. In contrast, the samples used in forensic DNA testing are often limited in quantity, degraded or contaminated, and come from unknown sources. Id. Thus, with little or no opportunity for repeat testing, accuracy is essential. Built-in consistency checks are lacking because testing often involves comparison of samples from many, not discrete, alternatives. In forensics, analysis of statistical population frequencies is extremely important in determining the significance of a match between samples.
An understanding of the procedures involved in DNA testing for forensic purposes is also crucial to any consideration of the admissibility of test results. We include a synopsis of this procedure to give a better understanding of the nature of the evidence the state seeks to admit.
DNA is the substance that carries the coded message of heredity and is often called the "blueprint of life." It is found in every cell that has a nucleus. DNA is the active substance of the genes which, in humans, is carried in 23 pairs of chromosomes. One chromosome of each pair is inherited from each parent. Every individual's overall DNA is distinct, except for identical twins, who have the same DNA.
The DNA structure is a double helix, made up of four chemical subunits commonly called bases. These bases pair predictably to form the rings of the double-stranded DNA helix. The unique sequence in the helix determines protein structure and the regulation of cell activities. Approximately 3.3 billion base pairs make up the entire genetic material contained in the chromosomes. Only a small number of these base pair sequences actually differ between any two individuals.
A DNA sequence has a specific physical location on a chromosome called a locus. Chromosomes contain many loci occupied by different DNA sequences. At each locus along the pairs of nonsex chromosomes, there usually are two alleles, or genetic variants, one inherited from each parent. However, the alleles may be identical, making the individual homozygous for that particular locus. Although an individual has no more than two alleles at a locus, many different alleles may exist within the human population for the same locus. When multiple alleles exist at a particular locus, the genetic variant is called a polymorphism. Detecting these differences is the key to DNA testing as it is used in forensics.
Forensic technology has recently been developed to remove DNA from human cells and to examine those loci where polymorphisms exist to determine whether a DNA sample, such as semen, found at a crime scene matches the suspect's DNA. Presently, only 3-5 loci are tested.
The DNA typing procedure used by the FBI is called restriction fragment length polymorphism (RFLP) analysis. It was first applied in forensic science in 1985 and is comprised of the following six steps:
(1) Extraction and Purification: Pure DNA is removed from an evidentiary sample of DNA using a chemical enzyme. Different testing laboratories use different restriction enzymes, which result in different DNA patterns for an individual.
(2) Fragmentation: Using restriction enzymes, which act like a pair of biological scissors, the DNA chain is cut into restriction fragments at specific loci. Each individual enzyme cuts at its own specific sequence whenever it is encountered along the DNA chain.
(3) Gel Electrophoresis: The cut fragments of DNA are separated by size within a gel. The DNA mixture is placed on the gel and exposed to an electric field. The negatively charged DNA fragments move toward the positive end of the gel. The larger fragments move more slowly than the smaller fragments so that the DNA mixture is separated according to the size of the fragments.
The FBI laboratory in Washington, D.C. incorporates ethidium bromide, a fluorescent dye that allows DNA to be visualized, into the gel before electrophoresis. Other laboratories stain the gels with ethidium bromide after electrophoresis. The National Research Council report recommends the latter method, because use of ethidium bromide prior to electrophoresis can alter the mobility of fragments if a high concentration of DNA is used in testing. See NRC Report at 12-7.
(4) Southern Blotting or Transfer: The DNA fragments are denatured and neutralized and then transferred from the gel to a nylon membrane, while retaining the same positions they previously occupied on the gel.
(5) Hybridization: The nylon membrane is placed in a bath that contains a single strand of a known DNA sequence, known as a single-locus probe. The probe is labeled with a radioactive isotope that facilitates the visualization. The probe locates and binds to a specific complementary fragment of DNA on the nylon membrane.
The probe binds to one or two fragments in each lane, depending on whether the individual is homozygous or heterozygous for that particular allele. The probe may also detect only one band in cases where the second allele is so small that it has migrated off the end of the gel, is not resolved, or is lost in partially degraded samples. Thus, an individual who is actually heterozygous with respect to a particular allele may appear to be homozygous.
(6) Autoradiography: The nylon membrane is placed between two pieces of high speed x-ray film called autoradiographs. It is placed in a freezer where the radioactive material begins to decay, exposing the x-ray film at the locations of the probe. At these locations, dark bands appear, leaving a DNA print or profile.
At this point, the physical process of RFLP analysis ends and the interpretation of the test begins. An analyst visually compares autorads from two separate DNA samples by "eyeballing" to determine if there is a match. A match occurs if the sizes and number of the detected DNA fragments, denoted by bands, are within a permissible degree of variance. If the visual inspection indicates a match, there is a computer measurement of the bands. The computer places a marker on each band — by locating the darkest sites on the auto-rad — and reports a numerical measurement of the band. The computer declares a match based on the same degree of variance, known as a match criteria or match window.
If the test does not exclude the suspect, then an analyst interprets the test results by estimating the probability that such a match would have occurred randomly. The analyst calculates the frequency in the population of each of the alleles that have been found, based on data bases that compile these frequencies in a previously studied group of individuals. Where there are two alleles, or bands, at a locus, the analyst may combine the individual band frequencies into a locus frequency.
I. Critical Impact
To prevail on a pretrial appeal, the state must show clearly and unequivocally that the district court erred in its judgment and that, unless reversed, the error will have a critical impact on the outcome of the prosecution. State v. Webber, 262 N.W.2d 157, 159 (Minn. 1977). "Critical impact" is shown "where the lack of the suppressed evidence significantly reduces the likelihood of a successful prosecution." Joon Kyu Kim, 398 N.W.2d at 551.
The state seeks to introduce limited statistical evidence regarding a matching DNA profile and nonstatistical opinion testimony concerning the significance of a match. With respect to the DNA test results, the district court ruled that experts would be limited to testimony that indicated the test results did not exclude Alt as a possible source of the body fluids tested. Statistical evidence of the likelihood of paternity was implicitly found to have critical impact in State v. Boyd, 331 N.W.2d 480 (Minn. 1983). The lack of the suppressed DNA testing evidence, either the statistical evidence or the nonstatistical opinion testimony, has critical impact because it significantly reduces the likelihood of a successful prosecution, particularly in light of the victim's apparent failure to identify Alt in a photo display. Moreover, Alt does not argue that critical impact is lacking. We conclude the district court's order does have critical impact.
II. Admissibility of DNA Test Results
Alt challenges the admissibility of the FBI DNA test results on a number of grounds, which fall into two broad categories: (1) challenges to the technical procedures themselves; and (2) challenges to the lack of proper "auditing" or other means of quality control.
The state argues that Alt's cross-appeal should be dismissed for failure to comply with the timing requirements for the filing of briefs under Minn.R.Crim.P. 28.04, subd. 2(3). However, that rule by its terms applies to appellant's briefs, and does not indicate that a respondent is to be treated as an appellant for purposes of his cross-appeal. Cross-appeals do not generate separate briefing deadlines; therefore, any deadlines to be imposed would not be jurisdictional. See Minn.R.Civ.App.P. 128.02, subd. 2 (respondent's brief to present issues raised in notice of review).
This court has held that Minn.R.Crim.P. 28.04, subd. 3 does not create an absolute right to review. State v. Joon Kyu Kim, 374 N.W.2d 814, 816 (Minn.App. 1985), aff'd on other grounds, 398 N.W.2d 544, 550 n. 8 (Minn. 1987) (noting district court has discretion to permit defense cross-appeal). However, judicial economy is served by considering all the DNA issues raised in a single appeal; therefore, we decline to dismiss the cross-appeal.
The supreme court has held:
[A]dmissibility of specific test results in a particular case hinges on the laboratory's compliance with appropriate standards and controls, and the availability of their testing data and results.
Schwartz, 447 N.W.2d at 428. The district court ruled admissible the FBI conclusion that Alt's DNA was consistent with the evidence sample. The court, however, excluded any evidence concerning the statistical frequency with which a random match could be expected. We discuss here only Alt's challenges that concern the basic conclusion of a DNA "match."
The state asks this court to rule on the continued necessity of a Frye hearing in individual cases.
The supreme court has held that laboratory DNA testing is not "so routine and the Schwartz guidelines so universally followed" that there is no need for a foundational Frye hearing when DNA test results are offered. State v. Nielsen, 467 N.W.2d 615, 619-20 (Minn. 1991); see also Jobe, 486 N.W.2d at 420 (Frye hearing still required because of evolving state of DNA forensic science).
As the state notes, the Frye hearing in this case is the third one concerning the FBI DNA procedures that has been reviewed in a published opinion. We conclude that, although some of the issues raised here have been raised before, at least in substantially the same form, a Frye hearing on FBI DNA procedures is still required under Nielsen and Jobe.
The Eighth Circuit Court of Appeals has generally approved pretrial admissibility hearings for DNA evidence even if most of the defense testimony would concern weight rather than admissibility. See United States v. Two Bulls, 918 F.2d 56, 61 (8th Cir. 1990) (the court should "hear both sides of the issue" in determining the scientific acceptability and reliability of testing procedures), reh'g granted, 925 F.2d 1127 (1991), reh'g vacated and appeal dismissed (Apr. 18, 1991).
The state rightly points out that there is substantial overlap between issues addressed at the Frye hearing and those presented at trial. However, one court has concluded:
It is the view of this court that given the complexity of the DNA multi-system identification tests and the powerful impact that they may have on a jury, passing muster under Frye alone is insufficient to place this type of evidence before a jury without a preliminary, critical examination of the actual testing procedures performed in a particular case.
People v. Castro, 144 Misc.2d 956, 545 N.Y.S.2d 985, 987 (N Y Sup. Ct. 1989).
To address this issue and those which follow, we must first discuss the recent Supreme Court opinion, Daubert v. Merrell Dow Pharmaceuticals, Inc., ___ U.S. ___, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), in which the Court held that the Frye test, "absent from and incompatible with the Federal Rules of Evidence, should not be applied in federal trials." Id. at ___, 113 S.Ct. at 2794.
Daubert does not by its terms apply to state court proceedings. Nonetheless, Daubert's persuasive force in Minnesota may be strengthened by the fact that our rules of evidence are modeled after the federal rules. See Minn.R.Evid. 101 cmt. The survival of Frye in Minnesota, however, has not been fully briefed or argued in this case. We note that our supreme court, in promulgating the Minnesota Rules of Evidence, expressly reserved the common law power to refine evidentiary standards through case law. Order Promulgating the Rules of Evidence (Minn. April 1, 1977), reprinted in 50 Minn.Stat.Ann. VII-VIII (West 1980). The supreme court has allowed Frye to flourish, even following promulgation of those rules. See, e.g., State v. Mack, 292 N.W.2d 764, 768 (Minn. 1980). It reaffirmed Frye as recently as the 1989 Schwartz opinion. Schwartz, 447 N.W.2d at 424-25. It is for our supreme court, not this court, to decide Daubert's impact in Minnesota. Cf. Iowa Nat'l Mut. Ins. Co. v. Auto-Owners Ins. Co., 371 N.W.2d 627, 629 (Minn.App. 1985) (supreme court is proper forum for addressing extension of existing law), pet. for rev. denied (Minn. Oct. 18, 1985).
Respondent moves to strike the state's letter argument submitted following release of Daubert. Since we find the state's argument unpersuasive, it is unnecessary to strike the submission, and we deny respondent's motion.
The supreme court has twice addressed the admissibility of FBI laboratory DNA test results. See State v. Johnson, 498 N.W.2d 10, 14 (Minn. 1993); Jobe, 486 N.W.2d at 419-20. In both cases, the court has concluded that the FBI DNA testing procedures meet the standards of reliability and admissibility set forth in Schwartz. However, Alt's challenges raise issues not addressed in Jobe and Johnson. For the following reasons, we conclude the FBI DNA procedures meet the Schwartz standards.
1. Technical challenges
Alt challenges the FBI's DNA technical procedures in two areas: (a) the procedures for declaring a "match," particularly the size of the match "window" and the allowance for analysts to override a computer "match" decision; and (b) the possibility of "band shifting," whether from the FBI's use of ethidium bromide or problems with the gel used in the gel electrophoresis stage.
a. Declaration of a "match"
Frye hearing testimony established that the FBI declares a match when the band length of the suspect's DNA sample for that allele is within 2.5%, plus or minus, of the forensic DNA sample. If the difference falls outside this match criterion, the FBI refers to the comparison as either inconclusive or as excluding the suspect. Alt offered testimony that this match "window" is too inclusive and could result in false positives.
The NRC Report does not specify what size match "window" is appropriate. The report recommends only that the match criterion be based on reproducibility studies showing actual degrees of observed variability in measurements of samples from the same person. NRC Report at 2-12. There was testimony that the FBI developed its match criterion in this way.
The supreme court in Jobe rejected a challenge to the FBI procedures which is very similar to that raised by Alt. Jobe, 486 N.W.2d at 419 (rejecting argument that FBI procedures for declaring a match lack an objective standard and allow differing conclusions by different analysts). Other courts have similarly rejected challenges to the FBI match criterion, or have held they go to the admissibility or weight of the DNA evidence in an individual case. See People v. Barney, 8 Cal.App.4th 798, 10 Cal.Rptr.2d 731, 739-40 (1992) (use of match criteria is part of inquiry on whether proper scientific procedures were used in the particular case); State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 492-93 (1992) (match "windows" and degree of objectivity go to admissibility or weight in individual case). The values reported in this case establish that any excess width in the "match window" did not affect the test results.
Four single locus probes were run on a specimen of Alt's DNA and three specimens recovered from the victim's clothing. The greatest variance between Alt's DNA and any of the forensic DNA specimens on any of the probes is 1.3%, approximately half the size of the match window, and within the match windows suggested by the defense experts.
An FBI analyst can manually override the computer's placement of the marker on a band. In placing the marker on the band to aid in determining the existence of a match, the computer is programmed to seek out the darkest portion of the band. Ideally, this will be the very center of the band. However, a band may appear much larger than it is, because of the excess amount of DNA, such that the computer's placement of the marker may not be in the center of the band. At this point, the analyst will use a manual override to place the marker at the center of the band. See generally, Lorne T. Kirby, DNA Fingerprinting: An Introduction 117-18 (Stockton Press 1990).
An FBI analyst testified that no record is kept of such an override, although other analysts later check the autoradiograpbs for accuracy. Alt challenges this lack of a reporting system on examiner overrides. As noted above, the match between Alt's DNA and the forensic samples is very close on all loci, well within the FBI's "matching window" of 2.5%. Alt has not shown that the declaration of a match could have been affected by a difference between the visual estimate and the computer sizing of the respective bands.
b. Band shifting
Alt challenges the admissibility of the DNA test results based on the possibility of "band shifting." There was testimony that the FBI use of ethidium bromide before electrophoresis may cause some band shifting, and that most DNA laboratories do not use ethidium bromide to stain the gel before electrophoresis. There also was testimony of other possible causes of band shifting.
"Band shifting" is a result of differences in movement between lanes of the gel during electrophoresis, resulting in altered band lengths. Office of Technology Assessment, Genetic Witness: Forensic Uses of DNA Tests 63 (1990).
The NRC Report acknowledges that DNA analysis is subject to band shifting. NRC Report at 2-4. The report concludes "there is no present justification for use of ethidium bromide in analytical gels." Id. at 2-7. The report also concludes that band shifting is easy to detect by using monomorphic probes. Id. at 2-10; see also Caldwell v. State, 260 Ga. 278, 393 S.E.2d 436, 442-43 (1990) (describing monomorphic probe test for band shifting). The FBI does not use monomorphic probes to detect band shifting. However, the NRC Report describes the possibility of an erroneous match due to band shifting as "rare." Id. at 2-11; see also Kirby, DNA Fingerprinting at 120 (although band shifting is a potential problem, it would result in a "no match").
The supreme court in Jobe urged that law enforcement agencies, prosecutors, and courts give consideration to the NRC Report's recommendation that monomorphic probes be used to detect band shifting. Jobe, 486 N.W.2d at 420 n. 3 ( Jobe was decided after the testing in this case). The court noted, however, that the appellant presented no evidence suggesting band shifting had occurred in that case. Id. at 420. Likewise, there is no evidence that band shifting affected the reliability of the DNA match in this case. Given the rarity with which band shifting would cause a false positive result, we agree with those courts that have held that this problem does not affect the basic reliability of a lab's DNA testing procedure, but relates only to the admissibility or weight of the results in an individual case. See, e.g., Barney, 10 Cal. Rptr.2d at 739-40.
Alt raises two other creditable issues regarding the test interpretation phase of the RFLP procedure.
The NRC Report recommends further testing to determine whether an observed one-banded pattern is truly a homozygote only in limited circumstances. NRC Report at 2-8. The FBI analyst in this case testified that there was no indication that a false one-banded pattern, detectable through further testing, existed. Alt has also failed to show that any failure to detect bands greater than 10,000 base pairs affected the results in this case.
2. Auditing issues
Alt challenges the quality control of FBI DNA testing on two principal grounds: (a) lack of blind proficiency testing, or testing by outside analysts; and (b) lack of general access of the scientific community to FBI data bases.
a. Proficiency testing
Frye hearing testimony established that the FBI does not test its DNA analysts by means of "blind" proficiency tests in which an analyst would be unaware of the test. Rather, the FBI uses "open" testing in which the analyst is given known samples, not disguised as forensic samples, to run through the RFLP test procedures. The NRC Report recommends blind proficiency testing and measurement of laboratory error rates with appropriate proficiency testing as part of a program of quality control. NRC Report at 3-25, 4-3.
The supreme court, in addressing a similar challenge to FBI DNA analysis, has noted the FBI's attempt to comply with the guidelines:
[T]here had been no independent audit of the FBI laboratory where the testing was done, because there was no one outside the FBI qualified to audit the facility. However, the FBI attempted compliance with section 10.1 of the guidelines by having the staff of its other DNA laboratory audit the laboratory in Washington, D.C.
Jobe, 486 N.W.2d at 419. The supreme court concluded the FBI DNA testing procedures are not inconsistent with the NRC Report:
There was testimony that the recommendation of outside proficiency testing is the only TWGDAM recommendation with which the FBI laboratory does not comply. There was also testimony that no agency outside the FBI was considered qualified to do such proficiency testing. This situation is unfortunate. See NRC Report at 2-5 ("there is no substitute for rigorous external proficiency testing via blind trials"). However, there appears to be some disagreement within the forensic science community about what kind of proficiency testing programs should be implemented, who should run them, and how the results should be used. Forensic Uses of DNA Tests at 79-80. The NRC Report recommends that laboratory error rates, as measured by proficiency tests, "should play a role in the interpretation of results of forensic DNA typing." NRC Report at 3-25. The Supreme Court's recent opinion abandoning Frye in the federal courts recognizes that "known or potential rate[s] of error" should play some role in the decision whether to admit scientific evidence. Daubert, ___ U.S. at ___, 113 S.Ct. at 2797.
This general problem of the lack of generally-accepted quality control guidelines is noted in Lisa Bouwer Hansen, Stemming the DNA Tide: A Case For Quality Control Guidelines, 16 Hamline L.Rev. 211, 230-31 (Fall 1992). However, the lack of quality control guidelines and external quality assurance programs is not limited to DNA testing, but is common in many areas of forensic science. Paul C. Giannelli, Criminal Discovery, Scientific Evidence, and DNA, 44 Vand. L.Rev. 791, 796 (1991).
While it is clear that a completely independent audit would be the ideal, there is no indication that the [FBI] test procedures used compromised the DNA testing results in this case.
Id. at 419-20. For similar reasons, we hold the DNA test results are admissible in this case.
b. Access to FBI data bases
Alt argues the FBI DNA test results are inadmissible because the FBI does not allow members of the scientific community general access to its data bases. The supreme court in Schwartz noted the importance of making a DNA laboratory's testing results and its data available to other scientists. See Schwartz, 447 N.W.2d at 428. The NRC Report recommends:
Any population databank used to support DNA typing should be openly available for scientific inspection by parties to a legal case and by the scientific community.
NRC Report at 3-25 (emphasis added).
Alt alleges specific examples of the FBI's denial of access by particular scientists. In responding to similar allegations of denial of access, the California Court of Appeals rejected a claim that the FBI has failed to share its data and methodology. See Barney, 10 Cal. Rptr.2d at 739. The court took judicial notice of "numerous published articles on DNA analysis as performed by the FBI." Id. Our supreme court has recently rejected, at least implicitly, the same claim. See Johnson, 498 N.W.2d at 14.
We are troubled by Alt's allegations of denial of access to the FBI data bases. However, neither this court nor the district court can possibly resolve all of the various disputes raised in the Frye hearing as to access, publication of critical articles, and other issues within the DNA forensic science community. Rather, we believe that although individual incidents may reflect on a laboratory's compliance with TWGDAM and the NRC Report access requirements, the proper focus in the Frye hearing should be on the policies and overall compliance of the laboratory. We hold, as the supreme court did in Johnson, that the DNA test results are not inadmissible on this basis.
In summary, we reject Alt's challenges to the admissibility of the DNA test results, but imply no limitation on the issues he may raise at trial. A defendant at trial may present evidence and arguments regarding the weight to be given DNA test results. Johnson, 498 N.W.2d at 14; Jobe, 486 N.W.2d at 420.
III. Statistical Frequency Evidence
The district court held that the FBI data bases were not generally accepted in the scientific community, and therefore ruled inadmissible any statistical frequency evidence.
The district court concluded that, in light of the debate over population subgroups, the FBI data bases did not meet the Frye test, and therefore the population frequency statistics could not be admitted at trial. The state noted the NRC Report's conservative approach to determining the population frequencies, called the "modified ceiling principle," and submitted an analyst's affidavit applying that principle in this case. The district court, while approving the NRC Report "modified ceiling principle," ruled initially that the state had not shown the FBI procedures met the requirements of the report. After giving the state an opportunity to provide more foundation, the district court confirmed its prior order.
The Frye test requires that a new scientific method from which proposed evidence is derived must be generally accepted by experts in the field as reliable and trustworthy. Schwartz, 447 N.W.2d at 424 (citing Mack, 292 N.W.2d at 768). Since 1990, all courts that addressed this issue noted the extensive debate among human geneticists and population geneticists concerning whether population subgroups make the estimates of the probability of a random match too low. See, e.g., Barney, 10 Cal.Rptr.2d at 740; United States v. Porter, 618 A.2d 629, 636-39 (D.C.App. 1992) (quoting Barney, 10 Cal.Rptr.2d at 740-43); Vandebogart, 616 A.2d at 493-94.
In calculating population frequencies based on its population data bases, the FBI has assumed that subgroups within populations do not exist. Dr. Richard C. Lewontin of Harvard University and Dr. Daniel Hartl of Washington University contend that estimated DNA frequencies are subject to potentially serious errors due to ethnic subgroups which may exhibit genetic differences due to nonrandom mating. Dr. Richard C. Lewontin Dr. Daniel L. Hartl, Population Genetics in Forensic DNA Typing, 254 Science 1745-50 (Dec. 20, 1991). Dr. Ranajit Chakraborty of the University of Texas and Dr. Kenneth K. Kidd of Yale University defend current DNA statistical interpretations, arguing that the range of possible error in estimated DNA frequencies is much smaller than suggested by Lewontin and Hartl. Dr. Ranajit Chakraborty and Dr. Kenneth K. Kidd, The Utility of DNA Typing in Forensic Work, 254 Science 1735-39 (Dec. 20, 1991). In summary, while some geneticists believe that the absence of subgroups and their effects on population frequencies cannot be assumed, others, while recognizing the possibility of the existence of subgroups, believe that they have minimal effect upon the calculation of population frequencies.
This debate affects principally the final step of DNA statistical analysis which is the application of the "product rule" to obtain the composite frequency of a match on all loci. The product of this step is not disclosed to the jury in Minnesota. See infra note 25. The possibility of subgroups does not affect only the final product, however. Subgrouping affects the presence of the condition of "Hardy-Weinberg equilibrium," an assumption made in calculating the frequency of a particular band in the population. Thus, the current scientific debate affects even the probability calculations on individual loci that are admissible in Minnesota.
Hardy-Weinberg equilibrium is the condition, for a particular genetic locus and particular population, with the following properties: allele frequencies at the locus are constant in the population over time, and there is no statistical correlation between the two alleles possessed by individuals in the population. Such a condition is approached in large randomly mating populations in the absence of selection, migration, and mutation. NRC Report at 8-5; see also Kirby, DNA Fingerprinting at 168-69; Castro, 545 N.Y.S.2d at 992-93.
The state argued at the Frye hearing that population substructure would affect only the final step, use of the "product rule" to yield the probability of a random match on all loci tested. However, the literature does not support this view. The bands visible at a particular locus could be inherited from either parent. Moreover, more than one band may be visible at a particular locus. Thus, the first step of the frequency calculation involves a multiplication of the frequency of each band (allele) at the locus. NRC Report at 3-5. This calculation assumes "that there is no statistical correlation between the allele inherited from one's father and the allele inherited from one's mother." Id.; see also Forensic Uses of DNA Tests, at 67 (statistical frequency calculation assumes there is no correlation at individual loci between alleles on maternal and paternal chromosomes). However, the possibility of population sub-grouping, because of nonrandom mating, calls into question this assumption. See Lewontin Hartl, Population Genetics, at 1746-47.
The population must be assumed to be in Hardy-Weinberg equilibrium for each locus in order for the initial individual-band frequency to be calculated. NRC Report at 3-5. The population subgroup debate challenges the assumption of Hardy-Weinberg equilibrium. As Lewontin and Hartl indicate, population sub-grouping affects whether reliable probability statements can be made concerning allele frequencies at individual loci. Lewontin Hard, Population Genetics, at 1747. Lewontin and Hard show a significant difference in the frequencies of alleles at locus D2S44, one of the loci tested here, between French and Israeli populations. Id. at 1748.
The NRC Committee has attempted to resolve this debate. The Committee's "modified ceiling principle" is a two-step process:
(1) For each allele at each locus, determine a ceiling frequency that is an upper bound for the allele frequency that is independent of the ethnic background of a subject; and (2) To calculate a genotype frequency, apply the multiplication rule, using the ceiling frequencies for the allele frequencies.
NRC Report at 3-10 to -11 (emphasis in original). In step one, the NRC Committee recommends a minimum interim ceiling frequency of 10%, i.e., no reported frequency of bands on an individual locus would ever be lower than 10%. The second step for the NRC method is irrelevant in Minnesota because the product rule would not be used.
Several courts have strongly suggested that statistical probability evidence as calculated by means of the NRC modified ceiling principle passes the Frye test and should be admitted. See, e.g., Barney, 10 Cal.Rptr.2d at 745 (NRC method appears to satisfy Frye requirement); Porter, 618 A.2d at 642-44 (noting scientific consensus satisfying Frye appears to exist on NRC approach, but remanding to trial court); Vandebogart, 616 A.2d at 494-95 (remanding issue of Frye consensus on NRC approach). We have found no case holding conclusively that this method satisfies the Frye test. Nevertheless, the Washington Supreme Court has observed:
Although we lack the scientific expertise to either assess or explain the methodology, its adoption by the [NRC] Committee indicates that sufficient acceptance within the scientific community has been achieved to satisfy Frye in appropriate circumstances.
State v. Cauthron, 120 Wn.2d 879, 846 P.2d 502, 517 (1993).
The district court re-opened the Frye hearing after the NRC Report was issued. At the hearing, respondent's counsel conceded the experts agree with the NRC Report, but claimed they do not agree with the FBI interpretation of it. The state agreed to stipulate to follow the NRC Report, and provided figures applying the modified ceiling principle to this case. The record fully reflects a basic agreement by the parties with the NRC modified ceiling principle, with the only disagreement concerning the FBI method of calculating the probabilities. Neither party has claimed the NRC's recommended calculation lacks acceptance within the scientific community.
This alleged disagreement concerned the multiplication rule and use of fixed bins, which do not affect the modified ceiling principle itself. The district court's exclusion of all statistical frequency calculations, including those admissible under Joon Kyu Kim and Johnson, was not based on a concern over the reliability of fixed bins, as opposed to floating bin calculations. The court was concerned about the population subgroup controversy and with its perception that jurors would multiply the individual band probabilities even without evidence allowing them to do so.
The court has discretion to exclude novel scientific evidence that does not meet the Frye test even though the parties have agreed to its admission. State v. Carlson, 369 N.W.2d 326, 328 (Minn.App. 1985) (exclusion of polygraph test results), pet. for rev. denied (Minn. July 26, 1985). However, the supreme court has held a defendant should be allowed to introduce unreliable scientific evidence (polygraph) where he seeks to do so to show the involuntary nature of his confession. State v. Schaeffer, 457 N.W.2d 194, 197 (Minn. 1990); see also State v. Kraushaar, 470 N.W.2d 509, 513 (Minn. 1991) (suggesting a defendant may waive Frye issue on appeal by failing to request a hearing).
We are reluctant to find a waiver of the Frye requirement as applied to the NRC recommendation. However, at issue is a conservative method of statistical calculation, designed to favor the defendant. We are not dealing with the admission of evidence, such as the polygraph test, whose reliability has been seriously questioned in numerous opinions. See, e.g., State v. Anderson, 379 N.W.2d 70, 79 (Minn. 1985), cert. denied, 476 U.S. 1141, 106 S.Ct. 2248, 90 L.Ed.2d 694 (1986).
Since the briefing in this case, such a dispute has arisen. See Peter Aldhous, Geneticists Attack NRC Report as Scientifically Flawed, 259 Science 755 (Feb. 5, 1993). However, this issue has not been litigated here. For a number of reasons, and particularly given the limited admissibility of statistical probabilities in Minnesota, we conclude the recent debate does not affect the admissibility in this case of DNA evidence under the NRC approach.
The NRC approach is a compromise position, designed to yield conservative estimates of random DNA matching which would be admissible in court. NRC Report at 3-23 to -24. The report incorporates a number of policy judgments, id., and does not purport to yield a statistical frequency which is more scientifically defensible than either the DNA proponents or opponents would advocate.
Insofar as the recent debate reflects disagreement with the policy judgments of the NRC Committee, or reiterates the view of DNA proponents that population subgroups are not a significant problem, the debate does not deprive the report of sufficient scientific acceptance under Frye. Without expert testimony in the record on this issue, we cannot assess to what extent this is a policy rather than a scientific debate. Parties in future cases may further litigate this issue.
A recent opinion views the debate more as a matter of policy:
While we are in no position to choose sides in this ongoing dispute, we note that its persistence threatens the admissibility of an extremely important forensic tool. This is no time for purist insistence that DNA evidence should be admitted on one's own terms or not at all.
People v. Wallace, 14 Cal.App.4th 651, 17 Cal.Rptr.2d 721, 726 (1993).
Although we have concluded the population subgroup issue does affect the limited statistics admissible in Minnesota, the impact is lessened because the individual loci frequencies are not multiplied. Although we believe the more recent debate over the NRC Report also is less significant here, parties in a future case will have an opportunity to fully litigate this issue.
We conclude the statistical frequencies of individual loci should be admitted in this case if calculated according to the NRC modified ceiling principle. Accordingly, we reverse the district court's decision that excluded this evidence. Parties in future cases may further litigate this issue. In this case, Alt may introduce evidence at trial which affects the weight to be given the FBI calculations under the NRC approach.
The district court also expressed the view that, if probability calculations were introduced for each loci, the jury would likely multiply those figures on its own, thus applying the "product rule" in a manner contrary to the NRC recommendation. The court held that because a jury might multiply the individual loci frequencies, those individual frequencies should not be admitted even if they were arrived at by a generally accepted scientific method. We disagree with the court's analysis. The court's complete exclusion of any individual loci frequencies on these grounds as stated in its order is contrary to Johnson and the rationale of Joon Kyu Kim.
IV. Nonstatistical Opinion Testimony
The state argues that the court erred in suppressing the nonstatistical opinion testimony of the significance of a DNA match. The district court has ruled that the state's DNA experts may testify only that the FBI DNA test does not exclude Alt as a possible source of the body fluids tested. The state asserts that the district court's ruling is contrary to the supreme court's holdings in Joon Kyu Kim and Jobe and to the general latitude given expert witnesses under Minn.R.Evid. 702.
The state's reliance on Jobe is misplaced. In Jobe, one of the state's DNA experts testified that the match of defendant's DNA with four samples recovered from the scene was a "very, very, very significant match." Jobe, 486 N.W.2d at 414. The supreme court in its analysis, however, discussed only the admissibility of the DNA "test results." Id. at 420. The court did not discuss the expert's nonstatistical description of the significance of the DNA match. Its holding that the test results were admissible does not imply approval of the expert's characterization of the significance of the match.
The supreme court, in addressing the admissibility of population frequency statistics for blood types, has held:
The expert should be permitted to testify, however, as to the basic theory underlying blood testing, to testify that not one of the individual tests excluded Kim as a source of the semen and to give the percentage of people in the general population with each of the individual blood types, and to express an opinion that scientific evidence is consistent with Kim having been the source of the semen.
Joon Kyu Kim, 398 N.W.2d at 549.
The state suggests its experts should be allowed to testify that Alt's DNA matches the forensic sample to a "reasonable degree of scientific certainty." However, the supreme court in Joon Kyu Kim carefully delineated the permissible bounds of expert opinion testimony in this area. Had the court intended to allow an expert opinion of a match to a "reasonable degree of scientific certainty," it would have done so explicitly. The court stopped short of allowing even the fact of a "match" to be stated in positive terms. See Joon Kyu Kim, 398 N.W.2d at 549 (expert may testify tests do not exclude defendant and are consistent with defendant being the source). Therefore, the state's proposed testimony exceeds the limits set in Joon Kyu Kim.
Another problem with the nonstatistical opinion testimony proposed by the state in this case is that it would not correctly reflect the nature of the DNA test results. Expert testimony is admissible if it will assist the trier of fact to understand the evidence or to determine a fact in issue. Minn.R.Evid. 702. The scientific community considers the statistical frequency important in interpreting DNA test results, but does not consider the opinion of a match to be significant. Nonstatistical opinion testimony on a DNA "match," without "the background probability information," is therefore not helpful to the jury. Cauthron, 846 P.2d at 516.
The problem is not that a probability of a random match which is, for example, 1 in 100,000, does not establish a "reasonable degree of scientific certainty." Rather, the problem is that the scientific community considers the statistical frequency important in interpreting DNA test results, and does not consider the opinion of a match to be significant. See Chakraborty Kidd, Utility of DNA Typing, at 1735 (DNA test results may be inconclusive, or may conclusively show no match; however, the significance of a possible match "should be evaluated in a legal setting").
The DNA expert's testimony would rely on the "product rule" multiplying the statistical probabilities of a match of the individual locus, which is not admissible in Minnesota. The probabilities of a random match within an individual locus are almost always high enough that an expert could not declare an overall match within a "reasonable degree of scientific certainty." See, e.g., Johnson, 498 N.W.2d at 14 (frequencies from 3 to 27 percent in different ethnic data bases).
Numerous courts have noted that the statistical frequency calculation is the essence of DNA evidence. One court has stated:
The statistical calculation step is the pivotal element of DNA analysis, for the evidence means nothing without a determination of the statistical significance of a match of DNA patterns.
Barney, 10 Cal.Rptr.2d at 742; see also Porter, 618 A.2d at 640 (probability calculation is at the core of DNA evidence).
We find some merit in the state's argument that without either the combined statistical probability figure or a nonstatistical opinion of the significance of a "match," it is deprived of the probative value of DNA evidence. As the Schwartz court stated, "Probability calculations are integral to DNA typing." Schwartz, 447 N.W.2d at 428. Indeed, DNA testing's inherent value is in the rarity of random matches, which gives DNA typing a discriminatory power far beyond that of conventional blood typing. See NRC Report at 3-16 (about 33% of suspects who match in conventional serology tests are excluded by DNA typing). Our supreme court, however, has consistently adhered to a position that statistical probability evidence, particularly evidence of the identity of the perpetrator of a crime, should be excluded because of its "potentially exaggerated impact on the trier of fact." State v. Carlson, 267 N.W.2d 170, 176 (Minn. 1978). The sentiment is echoed in the NRC Report, which recommends a cautious approach toward the use of random DNA match probabilities. We conclude the district court should have admitted testimony that DNA test results were consistent with Alt being the source of the forensic sample, as well as testimony that those results did not exclude him. However, we also conclude that further nonstatistical opinion testimony describing the significance of a match is barred by Joon Kyu Kim.
In Carlson, the court held that testimony of a 1 in 800 and a 1 in 4,500 chance that hairs found at the scene did not come from the defendant was erroneously admitted. Carlson, 267 N.W.2d at 176. In Boyd, the court held that expert testimony on the statistical probability that the defendant in a rape case was the father of the child born to the victim was properly excluded. Boyd, 331 N.W.2d at 483. The court noted the "real danger that the jury will use the evidence as a measure of the probability of the defendant's guilt or innocence," thus "erod[ing] the values served by the reasonable doubt standard." Id. The court held that the expert should be allowed to testify "as to the basic theory underlying blood testing" and that none "of the 15 tests excluded defendant as the father of the" child. Id. The court reiterated its opposition to population frequency statistics used to analyze blood test results in Joon Kyu Kim, 398 N.W.2d at 548. The court, however, allowed evidence as to the population frequency of the individual blood types, excluding only the frequency of the combination of blood types. Id. at 549. The court extended the Carlson-Boyd-Joon Kyu Kim analysis to DNA evidence in Schwartz, 447 N.W.2d at 428-29. The Joon Kyu Kim rule allowing population frequency evidence to this limited extent has been extended to individual DNA loci in Johnson, 498 N.W.2d at 15.
The issue of population subgroups is not the only problem with DNA statistical probabilities. Courts are admitting DNA test results from laboratories with no known or available error rates. One commentator has noted:
When the issue, however, is the trustworthiness of scientific evidence, courts generally cannot dismiss the possibility of error as purely theoretical or minimal.
Edward J. Imwinkelried, The Debate in the DNA Cases Over the Foundation for the Admission of Scientific Evidence: The Importance of Human Error as a Cause of Forensic Misanalysis, 69 Wn.U.L.Q. 19, 25 (1991). Another commentator suggests that the statistical probability of a coincidental match can never be lower than the rate at which the DNA testing laboratory errs in declaring matches (false positive rate). Richard Lempert, Some Caveats Concerning DNA As Criminal Identification Evidence: With Thanks to the Reverend Bayes, 13 Cardozo L.Rev. 303, 323-24 (1991).
A recent study suggests that, at least with respect to the interpretation of DNA samples, the false positive rate may be as low as nine cases in one million. I.W. Evett, DNA Statistics: Putting the Problems Into Perspective, 33 Jurimetrics J. 139, 143 (Fall 1992). However, another commentator concludes the false positive error rate is from one to four percent. Jonathan J. Koehler, DNA Matches and Statistics: Important Questions, Surprising Answers, 76 Judicature 222, 229 (Feb.-Mar. 1993).
Another problem in particular cases may arise from the high degree of genetic correlation among relatives. NRC Report at 3-14; see, e.g., Commonwealth v. Lanigan, 413 Mass. 154, 596 N.E.2d 311, 317 (1992) (DNA testing when two relatives were possible suspects).
DECISION
The district court properly admitted the FBI DNA test results. The court erred by excluding the FBI population frequency calculations on individual loci. The court properly limited expert nonstatistical opinion on the significance of a match, except testimony that the test results were consistent with Alt being the source of the forensic sample should be admitted.
Affirmed in part, reversed in part, and remanded.