Armstead v. State , 342 Md. 38 ( 1996 )

RAKER, Judge.

In this case we consider the statutory and constitutional limitations on the admissibility of deoxyribonucleic acid (DNA) evidence. Specifically, we must determine the effect of Maryland Code (1974, 1995 Repl.Vol., 1995 Cum.Supp.) § 10-915 of the Courts and Judicial Proceedings Article¹ on the admissibility of both the basic evidence of a DNA “match” and the descriptive statistics that are typically offered in support of a match. We shall hold that the statute renders both components of DNA evidence admissible. We shall also hold that the admission of DNA evidence in this case did not violate the Petitioner’s due process rights.

I.

On January 29, 1991, the victim, a Howard County woman, was at home alone when an assailant broke into the home, demanded her money and valuables, and then raped her and forced her to perform fellatio. The State presented evidence that pointed to the Petitioner, Michael Devon Armstead, as the perpetrator of these offenses. First, following the attack, the victim provided the police with a description matching Armstead. The victim also selected Armstead’s photograph from a photo array and identified him in court as the perpetrator. A neighbor who observed someone fleeing from the scene also identified Armstead as the perpetrator. In addition, when Armstead was arrested on the evening of the incident, he was wearing a leather jacket matching the victim’s description of the jacket worn by her attacker. The police also found a pair of pantyhose and a roll of duct tape in shrubbery near Armstead, which were both items that the perpetrator had used in the attack.

Physical evidence from the rape also linked Armstead to the crime. Semen was collected from the victim and analyzed using standard blood group testing. The blood group analysis indicated that Petitioner was within the 4.7% of the population that could have been the source of the semen. Finally, DNA analysis was performed using the restriction fragment length polymorphism (RFLP) testing method, revealing a “match” between the defendant’s blood and the semen sample taken from the victim.

Armstead was indicted in the Circuit Court for Howard County on twenty-five counts,² including charges of first and second degree rape, first and second degree sexual offense, perverted practices, assault, battery, burglary, robbery, and theft.

Prior to trial, Armstead filed a motion in limine to exclude the DNA evidence on both statutory and constitutional grounds. His statutory argument was that § 10-915 permits what he described as an “inverse Frye-Reed hearing.” He argued that, although the Legislature may have found RFLP testing reliable when it enacted § 10-915, this does not imply that the General Assembly intended the statute to preclude all future inquiry into the technique’s reliability. In light of recent scientific developments, Petitioner argued, the State should have been required to prove current general acceptance as a prerequisite to admission of the DNA evidence.

Petitioner also asserted several constitutional arguments. First, he contended that § 10-915 was unconstitutionally vague and overbroad because it did not provide any standards for DNA testing. Second, he argued that the use of DNA evidence violated his right to due process under the Fourteenth Amendment of the United States Constitution and Article 24 of the Maryland Declaration of Rights because “an individual [must] not suffer punitive action as a result of an inaccurate scientific procedure.” Higgs v. Wilson, 616 F.Supp. 226, 230 (W.D.Ky.1985), vacated and remanded on other grounds, 793 F.2d 1291 (6th Cir.1986), aff'd in part, vacated in part, and remanded in part on other grounds sub nom Higgs v. Bland, 888 F.2d 443 (6th Cir.1989). Third, Armstead argued that the statute denied him his right to confrontation under both the federal and state constitutions. Finally, he argued that the statute violated separation of powers because the Legislature had invaded the province of the judiciary by enacting an evidentiary rule.

The trial court rejected Petitioner’s statutory argument on the grounds that the statute precluded the trial court from holding a hearing on the reliability of DNA evidence. The court held a five-day evidentiary hearing, however, to address his constitutional challenges.³ At the hearing, much of Petitioner’s argument focused on his due process claim. He specifically challenged the method used to calculate the odds of a random or coincidental match between his DNA and the DNA taken from the victim. Petitioner contended that the DNA evidence should be excluded because this probability calculation was based on a purportedly outmoded method known as the “product rule” rather than the newer “ceiling principle” method, rendering the data so unreliable as to deny him due process.

The trial court denied the motion in limine, ruling that the evidence was admissible by statute, that the statutory conditions for admissibility had been satisfied, and that Petitioner’s constitutional arguments lacked merit. First, the court rejected Armstead’s “void for vagueness” argument, holding that the vagueness doctrine did not apply to an evidentiary statute such as § 10-915. Second, the court rejected the Petitioner’s due process argument because, after hearing extensive expert testimony, the court concluded that the testing procedures used by the laboratories in this case did not render the results so unreliable as to violate Armstead’s due process rights. Third, the court rejected Armstead’s confrontation argument, finding that Armstead exercised his right to cross-examine the State’s witnesses at the hearing, and that he would be afforded another opportunity to do so at trial. Finally, the court rejected the separation of powers argument, concluding that the Legislature possesses the authority to change the rules of evidence. The court therefore held that the DNA evidence would be admissible at trial.

Armstead was tried before a jury in the Circuit Court for Howard County. At trial, the State called several expert witnesses who presented both the product rule and ceiling principle calculations to the jury, explaining the rationale for each method. The witnesses explained that the product rule calculation yielded odds of a random match between Arm-stead’s DNA and the DNA recovered from the victim of one in 480 million, while the ceiling principle calculation yielded odds of a random match of one in 800,000. The jury was also informed of the laboratory error rates. Petitioner did not call any expert witnesses at trial to challenge the State’s DNA evidence; he did, however, cross-examine the State’s experts and elicited testimony regarding the controversy over the proper method of calculating match probabilities.

On December 9, 1992, the jury convicted Armstead of first degree rape, first degree sexual offense, perverted practices, assault, burglary, and attempted robbery. He was sentenced to two consecutive life terms plus twenty years. He noted a timely appeal to the Court of Special Appeals.

Armstead raised five issues before the Court of Special Appeals, but only one related to the admission of DNA evidence. He argued the trial court should not have admitted the DNA for two reasons. First, he claimed the trial court erred by not conducting a preliminary hearing, his proposed “inverse Frye-Reed hearing,” to determine whether the evidence was reliable. Second, he argued the use of outmoded methods of analysis rendered the DNA evidence so unreliable as to violate due process.

The Court of Special Appeals affirmed the convictions in an unreported opinion. The intermediate appellate court held that the trial court did not err in refusing to conduct a preliminary hearing on the RFLP technique, stating that while “[t]he reliability of the RFLP testing procedure is always open to attack ... the DNA profile’s admissibility is incontestable.” The Court of Special Appeals also held that the trial court did not err in refusing to exclude the DNA profile based on the statistical methods used because the evidence was admissible by statute. Finally, the intermediate appellate court held that Armstead’s due process rights were not violated. We granted certiorari to consider the important questions presented.

II.

Petitioner raises five issues before this Court. At the outset, we note that the Petitioner does not question the reliability of the general principles underlying DNA profiling. Petitioner first argues that, despite the enactment of § 10-915, trial judges retain discretion to exclude DNA evidence if its probative value is substantially outweighed by its prejudicial effect; he therefore contends that the trial court erred by refusing to engage in such balancing. Second, Petitioner asserts that § 10-915 does not preclude the possibility of a preliminary hearing to challenge the reliability of DNA evidence, and that such an “inverse Frye-Reed hearing” should be held if new evidence comes to light that calls the reliability of a previously accepted scientific technique into question. Third, Armstead argues that population genetics statistics must meet the “general acceptance” standard articulated in Reed v. State, 283 Md. 374, 389, 391 A.2d 364, 372 (1978),⁴ and that § 10-915 only established the admissibility of “raw” evidence of a DNA match. He concludes that if the Frye-Reed test had been applied to the statistical techniques used in this case, the product rule methodology would not have met the requisite “general acceptance” standard. Fourth, Arm-stead argues that because the laboratory error rate greatly exceeded the probability of error due to random DNA matching, the statistics on the odds of random matching were meaningless. He reasons that the DNA evidence was therefore incapable of “proving identity” as required by the statute. § 10-915(b).

Finally, in addition to these statutory arguments, Petitioner asserts a constitutional challenge. He argues the statistical evidence on the possibility of a random DNA match was so unreliable, due to the laboratory error rate, the allegedly improper probability calculations, and the use of improper laboratory procedures, that his due process rights were violated.

The State contends that § 10-915 eliminated the need for trial courts to engage in Frye-Reed analysis, and eliminated the discretion of trial courts to engage in a case-by-case balancing of probative value against prejudicial effect. The State also argues that the statute contemplated the admission of both the basic evidence of a DNA match and the supporting statistical evidence because the statistics provide necessary contextual information, and because the statute explicitly refers to “allele frequency” data. § 10-915(b)(2)(v). Finally, the State argues that Armstead’s due process arguments lack merit. The State contends that Petitioner’s challenges to the statistical methodology, the laboratory procedures, and the laboratory error rate go to the weight of the evidence rather than its admissibility.

We will address some of the Petitioner’s issues together, consolidating the arguments to three questions:

1. How does § 10-915 impact the “gatekeeping” function of the trial court in screening DNA evidence?

(a) May the trial court conduct an “inverse Frye-Reed hearing” if the opponent of DNA evidence challenges its reliability?

(b) May the trial court engage in a weighing exercise to determine if the probative value of DNA evidence is substantially outweighed by the prejudicial effect?

2. Does § 10-915 encompass population genetics statistics, in addition to the “raw” evidence of a DNA match?

3. Did the application of the product rule calculation, the rate of laboratory error, or the specific laboratory procedures used in this case render the resulting data so unreliable as to violate the Petitioner’s due process rights?

We address each of these issues in turn below.

III.

In order to understand the legal issues presented, some scientific explanation on DNA testing is helpful. This information has already been presented in several Maryland cases, see, e.g., Keirsey v. State, 106 Md.App. 551, 665 A.2d 700 (1995); Cobey v. State, 80 Md.App. 31, 559 A.2d 391 (1989), cert. denied, 317 Md. 542, 565 A.2d 670 (1989), as well as cases from many other jurisdictions.⁵ For that reason, the description that follows is abbreviated.

A. History of the Forensic Use of DNA Profile Evidence

DNA profiling has been used for forensic purposes for nearly a decade. It was first used in a criminal case in the United Kingdom in 1985, Office of Technology Assessment, Genetic Witness: Forensic Uses of DNA Testing 8 (1990), and was subsequently adopted by the FBI in 1988. NRC Report at S-l to S-2. Maryland’s first appellate case addressing DNA profiling evidence was Cobey v. State, 80 Md. App. 31, 559 A.2d 391 (1989), cert. denied, 317 Md. 542, 565 A.2d 670 (1989). By 1990, DNA profiling had been used in over ten thousand cases in the United States. H. Lee et al., DNA Typing in Forensic Science, 15 Am.J.Forensic Med. & Pathology 269, 270 (1994); see also R. Chakraborty & K. Kidd, The Utility of DNA Typing in Forensic Work, 254 Science 1735, 1735 (1991). Since the technique was first introduced, the overwhelming majority of state courts that have considered DNA evidence have found it admissible.⁶ See Developments in the Law: Confronting the New Challenges of Scientific Evidence, 108 Harv.L.Rev. 1481, 1558 (1995).⁷

B. The Science of DNA

Deoxyribonucleic acid or DNA is the genetic material that provides the instructions for all human characteristics, from eye color to height to blood type. P. Hartman & S. Susskind, Gene Action 2 (1965). Many types of cellular material carry DNA, including some types of blood cells, semen, and hair follicles. R. Lewontin & D. Hartl, Population Genetics in Forensic DNA Typing, 254 Science 1745, 1746 (1991). DNA is a “double helix” molecule, similar to a spiral staircase or a twisted rope ladder. Lee et al., supra, at 270. Each strand or “side” of the ladder is composed of four types of building blocks known as nucleotides, which can be connected in any order to form a DNA chain. Id. at 270-71. It is the sequence of the nucleotides that conveys the information, in effect “spelling out” the genetic instructions. G. Beadle & M. Beadle, The Language of Life 193-94 (1966).

A strand of DNA contains an estimated 50,000 to 100,000 genes, each of which directs the construction of a specific protein. Hartman & Susskind, supra, at 37. In addition to this “meaningful” DNA, the chain also includes “spacer” or “junk” DNA between the genes. The total amount of DNA composing all of an individual’s genetic information includes over three billion individual nucleotides, and a typical gene for an individual characteristic may be made up of tens of thousands of nucleotides. Lee et al., supra, at 270.

While each individual’s total DNA profile is unique, large segments of DNA are common to everyone. Out of the three billion nucleotides making up a complete DNA strand, there will likely be about three million differences in the DNA sequence between two randomly selected individuals. NRC Report, supra, at S-7. Many of these differences will be found in the “spacer” DNA areas, particularly in the number of times a spacer sequence is repeated. These highly variable areas in the DNA strand are known as VNTR’s, for “variable number of tandem repeats.” Typically, a VNTR will contain between twenty and one hundred repeats of the same nucleotide sequence. See Lee et al., supra, at 272 (Fig. 4); Lewontin & Hartl, supra, at 1745-46.

In criminal investigations, DNA profiling is typically used to compare a suspect’s DNA with a sample of DNA taken from the crime scene. DNA profiling does not compare every nucleotide of the suspect’s DNA with every nucleotide of the sample DNA, but rather compares the two at selected sites that are likely to vary from person to person. It is possible, however, that sections of DNA taken from different people will match. To avoid this type of “random matching” error, comparisons are made at multiple sites or loci along the DNA chain. Typically, laboratories analyze four or five loci in conducting DNA comparisons, reducing the probability of random matches across all loci to a low level. See L. Roberts, Fight Erupts Over DNA Fingerprinting, 254 Science 1721, 1721-22 (1991).⁸

C. The Method of DNA Analysis

The basic process of DNA analysis is the same whether it is used for diagnostic or forensic purposes. The most widely used technique at present is restriction fragment length polymorphism (RFLP) analysis.⁹ RFLP analysis involves three basic steps. First, a whole DNA strand is cut into smaller pieces using restriction enzymes, which are essentially chemical “scissors” designed to cut the DNA chain wherever a particular sequence of nucleotides is found. J. McKenna et al., Reference Guide on Forensic DNA Evidence, in Federal Judicial Center, Reference Manual on Scientific Evidence 282 (1994). The result is a mass of DNA fragments of varying sizes. Id. The second step is to separate these fragments according to their size. Lee et al., supra, at 271-73. This is accomplished by passing a current through a gel medium containing the DNA. The fragments are negatively charged, so they will migrate toward a positive electrode. Their progress toward the electrode will vary depending on their size, and thus the fragments will spread out across the gel. Id. (Fig. 5). Using a process known as Southern Blotting, these fragments are transferred from the gel to paper and washed with a radioactive material that attaches itself to the DNA fragments. Id. at 273. When the paper is placed against a sheet of film, the radioactive material exposes areas of the film, producing a discernible pattern of dark bands. This “picture” is known as an autoradiograph. Each band on the autoradiograph represents a fragment of DNA. McKenna et al., supra, at 283. Finally, these banding patterns can be used for identification by comparing the banding pattern in the suspect’s DNA with the pattern derived from DNA extracted from crime scene evidence. Id. at 283-84.

IY. Admissibility of DNA Match Evidence

A. General Principles of Statutory Construction

In Maryland, novel scientific evidence may become admissible in one of several ways. First, the evidence may be admitted by statute, if a relevant statute exists. See 5 L. McLain, Maryland Evidence § 401.4(c), at 277-78 (1987). Second, the proponent can prove that the evidence meets the Reed standard of “general acceptance” in the relevant scientific community. Reed v. State, 283 Md. 374, 381, 391 A.2d 364, 368 (1978) (quoting Frye v. United States, 293 F. 1013, 1014 (D.C.Cir.1923)). This can be accomplished through expert testimony, judicial notice, or a combination of the two. Goldstein v. State, 339 Md. 563, 567, 664 A.2d 375, 376-77 (1995). In the present case, the first method applies because the Legislature, by enacting § 10-915, declared DNA profiling evidence reliable and admissible.

Section 10-915 provides, in pertinent part,

(a) Definitions—

* # * * * ❖

(2) “Deoxyribonucleic acid (DNA)” means the molecules in all cellular forms that contain genetic information in a patterned chemical structure of each individual.

(3) “DNA profile” means an analysis that utilizes the restriction fragment length polymorphism analysis of DNA resulting in the identification of an individual’s patterned chemical structure of genetic information.

(b) Purposes.—In any criminal proceeding, the evidence of a DNA profile is admissible to prove or disprove the identity of any person....

The only condition the statute imposes on admission of DNA evidence relates to a discovery requirement, viz, information the proponent of the DNA evidence must provide to the opponent on request. §§ 10—915(b)(1)—(b)(2).¹⁰

The question we must consider is how to interpret the effect of § 10-915 on the traditional gatekeeping role of the trial court in determining the admissibility of DNA evidence. Petitioner raises two related questions in this regard, which we will analyze together. First, although Petitioner concedes that § 10-915 eliminates the need for a Frye-Reed hearing as a prerequisite to admission of DNA evidence, he contends that the Legislature merely intended to create a rebuttable presumption of admissibility. This interpretation, he argues, would allow the possibility for an opponent to challenge DNA evidence via an “inverse Frye-Reed ” proceeding, in which the opponent would bear the burden of showing the DNA evidence to be unreliable. Second, Armstead contends that despite enactment of § 10-915, the trial court retains its discretion to balance the probative value of DNA evidence against its prejudicial effect.

Both of these issues are essentially matters of statutory interpretation. When construing a statute, our governing principle must be the Legislature’s intent because, as we have consistently stated, the cardinal rule in statutory construction is to effectuate the Legislature’s broad goal or purpose. Gargliano v. State, 334 Md. 428, 435, 639 A.2d 675, 678 (1994). The primary source of legislative intent is the language of the statute itself. Rose v. Fox Pool, 335 Md. 351, 359, 643 A.2d 906, 910 (1994). In reading the language, we apply common sense to avoid illogical or unreasonable constructions, Frost v. State, 336 Md. 125, 137, 647 A.2d 106, 112 (1994), and we ascribe to words their common meanings, unless the Legislature intended otherwise. See Mustafa v. State, 323 Md. 65, 73, 591 A.2d 481, 485 (1991).

If the language alone does not provide sufficient information on the Legislature’s intent, then courts will look to other sources to discern the Legislature’s purpose. Gargliano, 334 Md. at 436, 639 A.2d at 678. Alternatively, if the language itself is clear and unambiguous and comports with the apparent purpose of the statute, there may be no need to consider other sources of information to glean the Legislature’s purpose. Jones v. State, 336 Md. 255, 261, 647 A.2d 1204, 1206-07 (1994). Because the meanings of even common words may be context-dependent, however, we often proceed to consider other “external manifestations of legislative intent,” Tidewater v. Mayor of Havre de Grace, 337 Md. 338, 347, 653 A.2d 468, 472 (1995), such as the amendment history of the statute, its relationship to prior and subsequent law, and its structure. Shah v. Howard County, 337 Md. 248, 255-57, 653 A.2d 425, 428-29 (1995); Kaczorowski v. City of Baltimore, 309 Md. 505, 515, 525 A.2d 628, 633 (1987).

B. Interpretation of § 10-915

1. Effect of the Statute on the Possibility of “Inverse Frye-Reed Hearings”

Applying the canons of statutory construction outlined above, we conclude that the notion of an “inverse Frye-Reed hearing” is inapposite when evidence is deemed admissible by statute. When the General Assembly has enacted legislation rendering evidence admissible, “the only way to contest the validity of the underlying principles involved would be to argue that the statutes violate one’s right to due process of the law.” L. McLain, Maryland Evidence § 401.4(c), at 278 (1987 & 1994 Cum.Supp.). See also J. Murphy, Maryland Evidence Handbook § 1406(C), at 733 (2d ed. 1993 & 1995 Cum.Supp.).

In reaching this conclusion, as we have indicated, our touchstone is the intent of the Legislature in enacting § 10-915. It is significant that the plain language of the statute explicitly states that DNA evidence “is admissible to prove ... identity,” § 10-915(b) (emphasis added), rather than using conditional language such as “may be admissible.”¹¹ The General Assembly’s choice of language alone, therefore, strongly suggests that the Legislature intended DNA profile evidence to be admitted without reevaluation of the technique’s general reliability.

We next consider whether this reading of the language corresponds to the apparent purpose of the statute. The legislative history clearly demonstrates that the primary reason the General Assembly enacted § 10-915 was to render DNA evidence admissible without Frye-Reed analysis in each case. When the DNA legislation was initially proposed, the Senate Judicial Proceedings Committee’s Report explicitly stated that “[t]he intent of the bill is to eliminate the necessity of holding a ‘Frye-Reed’ hearing to prove that the technique has gained general acceptance in the relevant scientific community.” Senate Judicial Proceedings Committee, Report on House Bill No. 711, at 2 (1989). Before § 10-915 was enacted, DNA profile evidence was admissible only if the technique satisfied the Frye-Reed “general acceptance” test. Reed, 283 Md. at 389, 391 A.2d at 372. See also Wheeler v. State, 88 Md.App. 512, 524, 596 A.2d 78, 84 (1991); Cobey, 80 Md.App. at 38, 559 A.2d at 392. At that time, the issue was likely to be relitigated in each case. When the General Assembly enacted § 10-915 in 1989, it clearly intended to streamline this process. See Senate Judicial Proceedings Committee, Report on House Bill No. 711, at 2 (1989).¹²

The 1991 amendment of the DNA statute also illustrates the Legislature’s confidence in the reliability of DNA evidence, because the amendment narrowed the potential arguments against admitting DNA evidence from general attacks on the methodology to specific attacks on the procedures used in the case at issue. The primary effect of the amendment was to expand the background information a proponent of DNA evidence would be required to give to the opponent. Senate Judicial Proceedings Committee, Report on House Bill No. 1150, at 2 (1991). Initially, the duty to disclose background information on DNA testing offered in evidence only applied to the State, but the bill expanded this duty to apply to both the State and the defendant. Id. The proponent of DNA evidence now must provide copies of the autoradiographs, laboratory protocols, and additional information relating to the laboratory’s statistical methods. § 10-915(b)(2)(i)-(v). The amendment also extended the notice requirement when DNA evidence is to be used from 15 days before trial to 45 days before trial. Senate Judicial Proceedings Committee, Report on House Bill No. 1150, at 2 (1991). In addition, courts are no longer required to ensure the presence of anyone in the chain of custody on demand.¹³ Moreover, the proponent of DNA evidence is no longer required to submit all reports generated about the DNA analysis, nor all the laboratory’s notes and photographs. Id. By providing the opponent with detailed, case-specific information on the DNA analysis and giving the opponent more time to evaluate the information before trial, the amendments also indicate the Legislature’s intent to establish the general reliability and admissibility of the evidence, permitting the opponent to attack the weight of the evidence through cross-examination. House Bill No. 1150; 1991 Maryland Laws ch. 631, at 3447-49 (1991) (codified as amended at § 10-915).

Finally, as further evidence of the Legislature’s intent, we turn to the preamble to the statute, which states that:

[M]eans of identifying that unique DNA structure have been refined far beyond any previous means of human tissue analysis, to a level of scientific accuracy that approaches an infinitesimal margin of error[.]

1989 Maryland Laws ch. 430, Preamble, at 2893 (1989) (emphasis added). This statement plainly illustrates the Legislature’s view that DNA evidence was sufficiently reliable to warrant elimination of the Frye-Reed hearings.¹⁴

We conclude that interpreting the statute to permit either traditional Frye-Reed hearings or the “inverse Frye-Reed hearings” proposed by the Petitioner would be contrary to the Legislature’s intent. As stated in the Fiscal Note to House Bill 711, under the bill, “an extensive hearing process would not occur and, therefore, court proceeding costs would decrease.” Division of Fiscal Research, Maryland General Assembly, Fiscal Note (Revised) on House Bill No. 711, at 1 (May 19, 1989). This statement is significant because it demonstrates that the Legislature intended to eliminate any extensive hearings, not merely to change the nature of the hearings by shifting the burden from the State to the defendant. The Legislature enacted § 10-915 to save time and money. Merely shifting the burden to defendants to prove DNA evidence unreliable, rather than requiring the State to prove it reliable, would not effectuate this purpose. Valuable resources and great time expenditures would still be required because both types of hearings are resource-intensive procedures which require costly and time-consuming expert testimony.¹⁵ This the Legislature sought to eliminate. We therefore conclude that the statute eliminates not only traditional Frye-Reed hearings, but also “inverse Frye-Reed hearings.” 2. The Effect of § 10-915 on the Discretion of the Trial Court to Weigh Probative Value Against Prejudicial Effect

Petitioner’s second argument, that the trial court retains discretion to balance the probative value of DNA evidence against its prejudicial effect is also without merit. As one court observed in United States v. Yee, 134 F.R.D. 161 (N.D.Ohio 1991), aff'd sub nom United States v. Bonds, 12 F.3d 540 (6th Cir.1993), the Frye test was designed to serve the same purpose as the trial judge’s discretionary balancing of probative value against prejudice:

The Frye doctrine developed ... out of the same concerns that led to the adoption of Rule 403 [providing the authority to weigh probative value against prejudicial effect]: namely, the concern that lay jurors might be misled by testimony that was unfairly prejudicial, confusing, or misleading.

Id. at 212 (citing United States v. Brown, 557 F.2d 541 (6th Cir.1977)). By enacting § 10-915 and thereby eliminating Frye-Reed hearings, the General Assembly legislatively determined that the probative value of DNA outweighs any prejudicial effect. The Legislature, in doing so, implicitly rejected Petitioner’s general arguments that DNA evidence is inherently prejudicial, the argument that the “aura of mystic infallibility” associated with DNA overshadows all other evidence, or the argument that highly technical evidence is “overwhelming” to juries. Except for constitutional challenges, therefore, generalized attacks on DNA testing are now precluded.

Petitioner also argues that use of the language “is admissible” in § 10-915 rather than mandatory language such as “shall be admitted” indicates that the Legislature intended to permit judges to continue to exercise discretion to exclude DNA evidence in some situations. We have considered the Legislature’s choice of language above, see supra Section IY.A, and we conclude that the words “is admissible” have eliminated the discretion of the trial court to weigh probative value against prejudicial effect.

Although we find that § 10-915 has eliminated some of the trial court’s gatekeeping responsibilities with regard to DNA evidence, we emphasize that trial courts still exercise an important function in determining whether DNA evidence is logically relevant to the case at hand. As we noted in Reed, there is an important distinction between the trial judge’s discretion to evaluate relevancy as opposed to reliability:

The question of the reliability of a scientific technique or process is unlike the question, for example, of the helpfulness of particular expert testimony to the trier of facts in a specific case. The answer to the question about the reliability of a scientific technique or process does not vary according to the circumstances of each case. It is therefore inappropriate to view the threshold question of reliability as a matter within each trial judge’s individual discretion.

Reed, 283 Md. at 381, 391 A.2d at 368. See also Haines v. Shanholtz, 57 Md.App. 92, 98, 468 A.2d 1365, 1369 (1984), cert. denied, 300 Md. 90, 475 A.2d 1201 (1984).¹⁶

While ordinarily DNA evidence will be admissible, the trial judge retains the discretion to exclude DNA evidence if errors in the laboratory procedures render it so unreliable that it would not be helpful to the trier of fact. See Jackson v. State, 92 Md.App. 304, 323, 608 A.2d 782, 791 (1992), cert. denied, 328 Md. 238, 614 A.2d 84 (1992). We recognize that courts in other jurisdictions have adopted differing views regarding whether challenges to the laboratory procedures used in a specific case go to the admissibility of DNA evidence or merely to its weight. Compare United States v. Martinez, 3 F.3d 1191, 1197-98 (8th Cir.1993), cert. denied, 510 U.S. 1062, 114 S.Ct. 734, 126 L.Ed.2d 697 (1994), with United States v. Jakobetz, 955 F.2d 786, 800 (2d Cir.1992), cert. denied, 506 U.S. 834, 113 S.Ct. 104, 121 L.Ed.2d 63 (1992). See also United States v. Two Bulls, 918 F.2d 56 (8th Cir. 1990), vacated and dismissed as moot, 925 F.2d 1127 (8th Cir.1991). Some courts, following the line of cases beginning with the trial court decision in People v. Castro, 144 Misc.2d 956, 545 N.Y.S.2d 985 (Bronx County Ct.1989), require adherence to accepted DNA protocols as a predicate for admissibility,¹⁷ see, e.g., State v. Houser, 241 Neb. 525, 490 N.W.2d 168, 181 (1992); Ex Parte Perry, 586 So.2d 242, 250 (Ala.1991), while others have concluded that case-specific errors in the laboratory procedures should ordinarily be evaluated by the factfinder in determining the weight of the evidence. See, e.g., United States v. Chischilly, 30 F.3d 1144, 1152-53 (9th Cir. 1994), cert. denied, — U.S. -, 115 S.Ct. 946, 130 L.Ed.2d 890 (1995); United States v. Bonds, 12 F.3d 540, 561 (6th Cir.1993); People v. Wesley, 83 N.Y.2d 417, 611 N.Y.S.2d 97, 104, 633 N.E.2d 451, 458 (1994); Fishback v. People, 851 P.2d 884, 893 (Colo.1993); State v. Cauthron, 120 Wash.2d 879, 846 P.2d 502, 507 (1993); People v. Mohit, 153 Misc.2d 22, 579 N.Y.S.2d 990, 992 (Westchester County Ct.1992). We believe the better approach is generally to treat individualized errors in application of the DNA technique as matters of weight, but to permit trial judges discretion to exclude DNA evidence if such errors were made in the course of testing that the evidence would not be helpful to the factfinder.¹⁸

[II] In order to resolve the issues presented in this case, we must first distinguish Petitioner’s general challenges to DNA testing from his particularized challenges to the procedures used in his case because, as we have indicated, his general challenges to the DNA testing methodology have been precluded by statute, while his specific challenges remain within the trial court’s discretion. We conclude that two of Petitioner’s contentions—his challenge regarding the use of the product rule and his challenge regarding the rate of laboratory error as compared to the odds of random matching—are general challenges, and as such, they are precluded.

A constitutional challenge to the statute or its application stands on a different footing from other generalized challenges to DNA evidence. The trial court always retains the authority to consider constitutional challenges to the statute or its application. We review the decision of Petitioner’s due process claim in Section VI, infra.

Petitioner attempts to raise one objection to the DNA testing methods as applied in his specific case: the presence of “shadow banding” in the autoradiographs.¹⁹ Shadow bands are “extra” bands that appear in one autoradiograph but not another, even though both are believed to be derived from the same source. Shadow bands may either indicate that the two DNA samples do not match, or they may be due laboratory error, such as excessive cutting action by the restriction enzymes, a phenomenon known as “star activity.” NRC Report, supra, at 2-9 to 2-10.

Although Petitioner raised the issue of shadow banding before the trial court as part of his due process challenge, however, he did not argue that the trial court retained its discretion under the statute to exclude the DNA evidence due to the shadow banding. Therefore we shall not reach the issue in this appeal. Md.Rule 8-131(a).

To summarize, we hold that § 10-915 precludes generalized challenges to the admissibility of DNA evidence, except for constitutional challenges. It therefore eliminates both traditional Frye-Reed or “inverse Frye-Reed ” hearings and individualized balancing of probative value against prejudicial effect. The statute does, however, permit case-specific challenges to the manner in which a particular test was conducted. Although these particularized challenges ordinarily will go to the weight of the evidence rather than its admissibility, the trial judge retains discretion to exclude evidence if it is so unreliable that it would not be helpful to the factfinder. See, e.g., Reed, 283 Md. at 389, 391 A.2d at 372 (“Testimony based on a technique which is found to have gained ‘general acceptance ... ’ may be admitted into evidence, but only if a trial judge also determines in the exercise of his discretion, as he must in all other instances of expert testimony, that the proposed testimony will be helpful to the jury, that the expert is properly qualified, etc.”).

Accordingly, we hold that the trial judge did not abuse his discretion in declining to conduct an “inverse Frye-Reed hearing” and in refusing to balance the probative value of DNA evidence against its prejudicial effect.

Y. Admissibility of Population Genetics Statistics

The next issue we must consider is whether population genetics statistics are admissible under § 10-915. Again, to resolve the legal issue, it is helpful to have a basic understanding of the science of population genetics.

A. The Scientific Debate Regarding the Use of Population Genetics Statistics

For each genetic characteristic, there may be two or more variations or forms of the controlling gene, which are called alleles. NRC Report, supra, at 1-3. Each parent contributes one copy of each gene, so every individual has two copies or alleles of each gene. Id. For two-allele genes, i.e., genes with only a “form A” and a “form B,” an individual may end up with one of three possible combinations: AA, AB, or BB. Each combination of alleles is known as a genotype. Id. at 1-5 (Fig. 1-3). RFLP analysis examines a number of different alleles from a single strand of DNA. Id.

The first step in statistically analyzing the results of RFLP testing is to determine the frequency of occurrence of each allele tested in the general population. Id. at 3-2. In the case of a two-allele gene, form A may occur in 30% of the population, while form B occurs in 70% of the population. Therefore, the fact that an individual has form A of the allele is not, in itself, very informative because there is a 30% chance that form A would be found in a random member of the population.

Let us assume, however, that three different loci are tested. The alleles for each locus are Al and B1, A2 and B2, and A3 and B3, respectively. Further assume that the allele frequencies for the general population are 10% for all the “A” alleles, and 90% for all the “B” alleles. If the suspect’s DNA is found to include alleles Al, A2, and A3, the probability of a random match with this profile can be calculated by multiplying the probability of a random individual having allele A1 times the probability of having A2 times the probability of having A3, or 10% times 10% times 10%, which equals 0.1%.²⁰ See R. Lempert, The Suspect Population and DNA Identification, 34 Jurimetrics J. 1, 1-2 n. 3 (1993). By increasing the number of loci tested, this probability of random matching can be reduced further, so that if five alleles were tested, the probability of a random match would be only 0.001%, or one in one hundred thousand. See supra note 20. In actual practice, the probability of random matching is reduced even further by choosing highly variable areas of the DNA with dozens of different alleles, so that individual allele frequencies will be very low. Chakraborty & Kidd, supra, at 1735, 1736.

There are at least two significant potential problems in calculating the probability of a random DNA match as outlined above. The first question is how to combine the probability of random matching for each allele to come up with an overall probability of random matching across all alleles. The second question is how to select the proper reference group to be used to calculate the general allele frequency in the population. The essence of this query is whether the general population may be used, or whether there is enough difference in allele frequency across racial and ethnic population subgroups to require more specific subpopulation frequencies to be used.²¹ We note that although there was significant debate across the country in both the scientific and legal communities concerning these issues, compare Lewontin & Hartl, supra, with Chakraborty & Kidd, supra, this controversy has largely been resolved by recent scientific studies. See, e.g., B. Budowle et al., The Assessment of Frequency Estimates of Hae III-Generated VNTR Profiles in Various Reference Databases, J. Forensic Sci. 319, 349 (1994); U.S. Dep’t of Justice, VNTR Population Data: A Worldwide Study (1993); B. Devlin & N. Risch, A Note on Hardy-Weinberg Equilibrium of VNTR Data by Using the Federal Bureau of Investigation’s Fixed-Bin Method, 51 Am.J.Hum.Genetics 549 (1992); B. Devlin & N. Risch, Ethnic Differentiation at VNTR Loci, with Special Reference to Forensic Applications, 51 Am.J.Hum.Genetics 534, 545-47 (1992).

Most of the controversy over use of DNA evidence has focused on the first question, i.e., how to combine the probabilities of random matching across all alleles. The hypothetical allele frequency calculations described above rely on a probability principle known as the “product rule.” Stated generally, the product rule means that the probability of two events occurring together is equal to the probability that event one will occur multiplied by the probability that event two will occur. R. Freund & W. Wilson, Statistical Methods 62 (1993). The classic illustration is coin tossing; the probability of finding “heads” on two successive coin tosses is equal to the probability of heads on the first toss, 50%, times the probability of heads on the second toss, 50%, equalling 25%. R. Johnson, Elementary Statistics 143 (4th ed. 1984).

The product rule is valid if the individual events are independent, i.e., if the outcome of the first event does not impact the outcome of the second event. Freund & Wilson, supra, at 62. In the coin toss example, this means that the outcome of the first coin toss does not affect the outcome of the second coin toss, which is a valid assumption. Id. By comparison, assume we wish to calculate the probability of having both a checking account and a loan from a particular bank. This is an example of non-independent or linked events. Johnson, supra, at 144. We can not calculate the probability of having both a loan and a checking account at the same bank by multiplying together the individual probabilities under the product rule because a person is more likely to obtain a loan from the bank where he maintains a checking account. Id. To illustrate nonindependence as it applies to human characteristics (although not genetic characteristics), assume we wish to determine the probability a man will have both a beard and a moustache. Also assume that the probability of having a beard is 1/20, and the probability of having a moustache is 1/10. It would be incorrect to infer that the probability of having both a beard and a moustache, applying the product rule, is 1/200, because it is likely that these are non-independent events; men who have beards are probably more likely than others to also have moustaches. See People v. Collins, 68 Cal.2d 319, 66 Cal.Rptr. 497, 503 & n. 15, 438 P.2d 33, 39 & n. 15 (1968).²²

The legal and scientific debate regarding DNA evidence mainly revolved around whether or not the product rule could be applied to genetic testing. L. Roberts, Fight Erupts Over DNA Fingerprinting, 254 Science 1721, 1723 (1991); see also J. McKenna et al., Reference Guide on Forensic DNA Evidence, in Reference Manual on Scientific Evidence 300 (1994). In the past, population geneticists and other scientists disagreed over whether the probability of possessing one allele was really independent from the probability of possessing another allele, and over the importance of any such differences. Compare Chakraborty & K. Kidd, supra, with Lewontin & Hartl, supra. Theoretically, truly independent distribution of individual alleles in the population requires completely random mating across racial and ethnic divisions, under a genetic principle known as Hardy-Weinberg equilibrium.²³ Lewontin & Hartl, supra, at 1746-47; Chakraborty & Kidd, supra, at 1736. Since it is generally acknowledged that the population has not reached such a state of equilibrium at present, some scientists hypothesized that there might be significant substructuring in the population which could cause considerable variation in the allele frequency across subpopulations. Lewontin & Hartl, supra, at 1747; L. Mueller, Population Genetics of Hypervariable Human DNA, in Forensic DNA Technology 60 (1992). There was, however, no empirical data to support this theory. Chakraborty & Kidd, supra, at 1737-38.

In 1992, the National Research Council issued a report on forensic DNA testing to attempt to resolve this dispute. NRC Report, supra. Although the report recommended using a conservative modification of the product rule known as the ceiling principle, it did not ultimately reject the product rule. Id. at S-ll. Instead, it merely suggested that until data could be collected to confirm or refute the existence of significant population substructuring effects, the interim approach should be to incorporate several conservative assumptions into the product rule calculation. Id. at S-11 to S-12. The effect of these conservative assumptions is to maximize the likelihood of random matches, thus decreasing the power of the DNA results to some extent.²⁴

While the NRC report did not definitively resolve the issue, however, the debate over the product rule essentially ended in 1993, with the announcement in the scientific journal Nature that the “DNA fingerprinting wars are over.” E. Lander & B. Budowle, DNA Fingerprinting Dispute Laid to Rest, 371 Nature 735, 735 (1994). Eric Lander, formerly a vociferous opponent of use of the product rule, was one of the authors heralding this shift in scientific opinion.²⁵ Id. While a small number of scientists still advocate very conservative treatment of DNA analysis until more data is collected, see, e.g., More on DNA Typing Dispute, 373 Nature 98-99 (1995), the empirical data collected thus far has indicated that the population substructuring that currently exists does not result in forensically significant variation in allele frequencies across population subgroups.²⁶ See Lander & Budowle, supra, at 736; see also B. Devlin & N. Risch, A Note on Hardy-Weinberg Equilibrium of VNTR Data by Using the Federal Bureau of Investigation’s Fixed-Bin Method, 51 Am.J.Human Genetics 549 (1992); B. Budowle et al., The Assessment of Frequency Estimates of Hae III-Generated VNTR Profiles in Various Reference Databases, 39 J.Forensic Sci. 319 (1994). Ultimately, the studies have generally concluded that use of the ceiling principle is unnecessary:

[T]he data do not support the need for alternate procedures, such as the ceiling principle approach (NRC Report 1992), for deriving statistical estimates of DNA profile frequencies.

Estimates of the likelihood of occurrence of a DNA profile using each of the major population group databases (e.g., Caucasian and Black) provide a greater range of frequencies than would estimates from subgroups of a major population category. Comparisons across major population groups provide reasonable, reliable, and meaningful estimates of DNA profile frequencies without forensically significant consequences.

VNTR Population Data Study, supra, at 6 (emphasis added). Similarly, another study concluded that:

Subdivision, either by ethnic group or by U.S. geographic region, within a major population group does not substantially affect forensic estimates of the likelihood of occurrence of a DNA profile---- Estimated frequencies among regional groups and several subgroups of a major population category are similar.... The most appropriate approach, therefore, is to estimate the likelihood of occurrence of a particular DNA profile in each major group.... [Bjased on empirical data, there is no demonstrable need for employing alternative approaches, such as the ceiling principle, to derive statistical estimates. VNTR frequency data from major population groups provide valid estimates of DNA profile frequencies without significant consequences for forensic inferences.

Budowle et al., supra, at 349 (emphasis added). In recent cases, courts considering the admissibility of DNA evidence have cited the growing list of scientific publications refuting the Hartl and Lewontin theory and supporting the continuing validity of the product rule. See, e.g., People v. Smith, 42 Cal.App.4th 204, 49 Cal.Rptr.2d 608, 613-15 & n. 15 (1996); People v. Marlow, 39 Cal.App.4th 343, 41 Cal.Rptr.2d 5, 32-33 (1995), cert. granted, — Cal.3d -, 43 Cal.Rptr.2d 679, 899 P.2d 65 (1995); People v. Soto, 39 Cal.App.4th 757, 35 Cal. Rptr.2d 846, 855-56 (1994), cert. granted, — Cal.3d -, 39 Cal.Rptr.2d 406, 890 P.2d 1115 (Cal.1995).²⁷

In addition to the controversy between the product rule and the ceiling principle, however, scientists have also pointed to problems in the selection of a reference database. Lewontin & Hartl, supra, at 1746. See also Com. v. Curnin, 409 Mass. 218, 565 N.E.2d 440, 444 & n. 11 (1991). In the hypothetical calculation described above, see supra pages 66-67, we assumed that form A of the gene occurred in 30% of the population, while form B of the gene occurred in 70% of the population. In reality, however, it is difficult to determine the population frequency of a given allele. First, the relevant population must be defined. Lewontin & Hartl, supra, at 1746. If allele frequencies vary substantially across ethnic and racial population subgroups, then it may be necessary to base the population frequency on a subgroup corresponding to the suspect’s ethnic or racial category. Id. at 1747-49. If the suspect is Hispanic, therefore, the most conservative way to proceed is to use a Hispanic population group to determine the allele frequency.²⁸ Even within the Hispanic category, however, there is substantial potential for variation in allele frequencies, i.e., from Hispanics who are mainly of Indian descent versus Hispanics mainly of European descent. Id. at 1749 (“Because of the extreme heterogeneity among ‘Hispanics’ and among ‘native Americans,’ it is doubtful whether any reference population could be defined that would be reliable in a forensic context.”). See also Lempert, supra, at 2.

Further complicating the issue, even assuming one can determine the proper population database to draw from, the allele frequency must still be determined. This requires obtaining sample DNA profiles from some significant number of people, ranging from several hundred to several thousand. Scientists disagree over the appropriate minimum number of profiles that should be used to make a database meaningful. See Devlin et al., supra, at 749; see also People v. Soto, 39 Cal.App.4th 757, 35 Cal.Rptr.2d 846, 852 n. 15 (1994), cert. granted, 39 Cal.Rptr.2d 406, 890 P.2d 1115 (1995). Moreover, the more refined the subgroup analysis becomes, the smaller the subset of the total database that can be used, exacerbating the database size problem. As a result, if a database of several thousand profiles is limited, for statistical analysis purposes, to only those profiles belonging to Indian-descended Hispanics, this may reduce the reference database to only a few dozen profiles.²⁹

Resolution of the product rule versus ceiling principle debate has, however, also ameliorated the database selection problem. Since the majority of scientists now believe that the effects of population substructuring are relatively insignificant, it has become unnecessary to develop data for very small population subgroups. See generally VNTR Population Data Study, supra; E. Lander & B. Budowle, supra; Chakraborty et al., supra; Budowle et al., supra. See also Lempert, supra, at 3.³⁰

B. The Effect of Section 10-915 on Admission of Population Genetics Statistics

Although § 10-915 does not explicitly discuss the use of population genetics statistics, there are several indications in the statute that the Legislature also intended the supporting statistics to be routinely admitted along with the DNA match evidence. While the scientific dispute discussed above might at one point have required exclusion of population genetics under the Frye-Reed standard, the statute renders Frye-Reed analysis unnecessary.

First addressing the language of the statute, the General Assembly amended the DNA statute in 1991 to specifically include a provision referring to population genetics. 1991 Maryland Laws ch. 631, at 3447-49 (1991) (codified as amended at § 10-915); see also House Bill No. 1150. As amended, the statute now requires that the proponent of DNA evidence provide, upon the opponent’s request, “a statement setting forth the allele frequency and genotype data for the appropriate data base utilized.” § 10—915(b)(2)(v). See also supra note 10. This language clearly indicates that the Legislature was aware that population genetics were used in support of DNA evidence, and moreover, that the Legislature expected such information to be presented at trial and used in cross-examination. This amendment preceded Armstead’s trial, and Armstead received all the discovery information required under the amended statute.

In addition, the preamble language discussed above referring to an “infinitesimal margin of error,” also demonstrates that the Legislature knew that statistical calculations were routinely applied to gauge the accuracy of DNA profile evidence. See supra Section IV.B.l. Furthermore, the fact that the Legislature deleted the words “unique” and “uniquely” from House Bill 711 before enacting the DNA statute reflects an awareness that not all segments of DNA are unique and, therefore, that there is some possibility of random matching. Even before the statute was amended, therefore, the Legislature clearly recognized that the odds of random matching would be at issue whenever DNA evidence was presented.

Moreover, the statutory language stating that DNA profile evidence is admissible “to prove or disprove ... identity,” § 10-915(b), also indicates that the Legislature viewed population genetics statistics as a necessary component of DNA evidence. As stated in the NRC Report, “[t]o say that two patterns match, without providing any scientifically valid estimate (or, at least, an upper bound) of the frequency with which such matches might occur by chance, is meaningless.” NRC Report, supra, at 3-1. Courts in other jurisdictions have also recognized this problem, observing that: “Without the probability assessment, the jury does not know [whether the matching] patterns are as common as pictures with two eyes, or as unique as the Mona Lisa.” United States v. Yee, 134 F.R.D. 161, 181 (N.D.Ohio 1991), aff'd sub nom United States v. Bonds, 12 F.3d 540 (6th Cir.1993). See also Nelson v. State, 628 A.2d 69, 75 (Del.1993); State v. Cauthron, 120 Wash.2d 879, 846 P.2d 502, 516 (1993); United States v. Porter, 618 A.2d 629, 640 (D.C.1992); Com. v. Lanigan, 413 Mass. 154, 596 N.E.2d 311, 314 (1992) (Lanigan I), aff'd in part and rev’d in part, 419 Mass. 15, 641 N.E.2d 1342 (1994) (Lanigan II); State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 494 (1992), modified on reh’g, 139 N.H. 145, 652 A.2d 671 (1994); People v. Marlow, 39 Cal.App.4th 343, 41 Cal.Rptr.2d 5, 29 & n. 42 (1995), cert. granted, 43 Cal.Rptr.2d 679, 899 P.2d 65 (1995); People v. Soto, 39 Cal.App.4th 757, 35 Cal.Rptr.2d 846, 855 (1994), cert. granted, 39 Cal.Rptr.2d 406, 890 P.2d 1115 (1995); State v. Watson, 257 Ill.App.3d 915, 196 Ill.Dec. 89, 99, 629 N.E.2d 634, 644 (App.Ct.1994), cert. denied, 157 Ill.2d 519, 205 Ill.Dec. 182, 642 N.E.2d 1299 (1994).

If random DNA matching is possible, then a “match” between two DNA profiles is not meaningful without contextual statistics regarding the odds that the match was coincidental. United States v. Yee, 134 F.R.D. 161, 181 (N.D.Ohio 1991), aff'd sub nom United States v. Bonds, 12 F.3d 540 (6th Cir.1993); Lanigan, 596 N.E.2d at 314, Cauthron, 846 P.2d at 516. The General Assembly recognized the possibility of random matching, as indicated by the 1991 amendment and the preamble to the statute; therefore, in rendering DNA evidence admissible, we conclude that the Legislature intended to render the necessary contextual statistics admissible, not just the “raw” evidence of a DNA match. This interpretation is in accord with common sense because to interpret the statute otherwise would provide juries with DNA evidence that they could not evaluate in a logical manner.

We recognize that some courts have allowed use of DNA match evidence without supporting statistics.³¹ See, e.g., Com. v. Crews, 536 Pa. 508, 640 A.2d 395 (1994). We believe, however, that the better approach is to treat the match and the statistics as inseparable components of DNA evidence.³² Cf. Keirsey v. State, 106 Md.App. 551, 575, 665 A.2d 700, 712 (1995).³³

Thus, we conclude that § 10-915 encompasses both the evidence of a DNA match and the supporting statistics. Although § 10-915 does not specify which statistical methodology should be used, ie., the product rule or the ceiling principle, the Legislature need not reach this level of specificity in order for the statistical information to be admitted. For example, as originally enacted, the statute did not specify which type of DNA analysis was admissible; the Legislature only added the requirement that the evidence be collected using the RFLP method when the statute was amended in 1991. 1991 Maryland Laws ch. 631, at 3448 (codified as amended at § 10—915(b)). This lack of specificity can not be read to mean that DNA analysis was inadmissible until the Legislature committed to one specific method.

Moreover, if the Legislature specified the precise method to be used for each critical step in DNA analysis, it would stultify scientific development.³⁴ In Haines v. Shanholtz, 57 Md.App. 92, 468 A.2d 1365 (1984), cert. denied, 300 Md. 90, 475 A.2d 1201 (1984), the Court of Special Appeals discussed this point in describing why the Legislature chose not to codify the specific calculations to be used in determining the probability of paternity:

The legislation ... carefully refrained from adopting any specific tests to establish the percentages necessary to include or exclude putative fathers. It is reasonable to assume that the General Assembly recognized that new technology may become available and, by not addressing any specific combination of tests, laboratories will be allowed to utilize the most effective tests without further legislative change.

Id. at 96, 468 A.2d at 1367. To codify every step in a technological process such as DNA profiling would be to freeze the process in time, precluding the introduction of better procedures and potentially more accurate evidence without a subsequent change in the law. Cf. State v. Bible, 175 Ariz. 549, 858 P.2d 1152, 1189 n. 33 (1993), cert. denied, - U.S. -, 114 S.Ct. 1578, 128 L.Ed.2d 221 (1994).³⁵

Petitioner is simply wrong when he argues that the Court is obliged to reconsider the reliability of the product rule in light of the NRC Report and other recent scientific developments. While due process considerations require courts to intervene if scientific opinion shifts so dramatically that previously accepted methods are considered unreliable, § 10-915 shifts the responsibility to the Legislature to respond to incremental advances in technology that do not effect sufficient change to implicate due process. Although Petitioner argues, as discussed in Section VI, infra, that there has been a significant shift in scientific opinion, such that the product rule has become outmoded, we disagree. Neither the NRC report nor the ensuing scientific publications support such a conclusion. See NRC Report, supra, at S-11 to S—12; B. Budowle & E. Lander, DNA Fingerprinting Dispute Laid to Rest, 371 Nature 735, 737-38 (1994). On the contrary, both methods have been proven viable in light of the most recent scientific evidence. See VNTR Population Data Study, supra, at 6; Budowle et al., supra, at 349. Section 10-915 does not specify a particular statistical method; therefore, it was appropriate for the trial court to permit the results of both calculations to be presented in court. Cf. Soto, 35 Cal.Rptr.2d at 858-59 (affirming the trial court’s decision to admit both the product rule and ceiling principle calculations under the Kelly-Frye standard).

We hold that the trial court did not abuse its discretion in admitting the statistical evidence regarding the probability of a random DNA match calculated using the product rule because the statute contemplates the use of probability evidence to place the “raw” result of a DNA match in context. Furthermore, the Legislature need not specify the particular statistical method to be used. Because the statute does not specify a method, either the product rule or the ceiling principle may be applied and presented in evidence.

VI. Due Process Considerations

Petitioner next contends that the DNA evidence is so unreliable that its use violates his due process rights. He bases his contention that the DNA evidence was unreliable on both generalized problems with DNA testing and specific problems with the conduct of the testing in his case. His generalized attack focuses on the use of the product rule rather than the ceiling principle, which he contends renders the statistical component of the DNA evidence unreliable. He also argues that because the laboratory error rate, 0.7%, greatly exceeds the odds of random matching under either the product rule or the ceiling principle, both methods of calculating the odds are unreliable.

In addition, he attacks several specific defects in the conduct of the testing in his case. The first technical defect he discusses is the use of excessively large “match windows” in comparing his DNA with the DNA taken from the victim. Match windows must be used due to measurement error that is inherent in the DNA technique. Even if an identical piece of DNA was measured several times, the measurement would likely differ to a small degree. Therefore, match windows are used to account for this margin of error.³⁶ If match windows are too large, two fragments may be declared a match when they actually differ. Petitioner argues that this problem rendered his test results unreliable. Petitioner also argues that error in the testing yielded bands in one test that did not appear in another test, a phenomenon the laboratories attributed to “star activity,” or essentially excessive activity of the restriction enzymes used to cut the DNA. Third, he argues that the statistics on odds of a random match failed to account for the higher degree of genetic similarity between Petitioner and his siblings.

We begin our analysis with the proposition that “a part of the due process guarantee is that an individual not suffer punitive action as a result of an inaccurate scientific procedure.” Higgs v. Wilson, 616 F.Supp. 226, 230 (W.D.Ky. 1985), vacated and remanded on other grounds, 793 F.2d 1291 (6th Cir.1986), aff'd in part, vacated in part, and remanded in part on other grounds sub nom Higgs v. Bland, 888 F.2d 443 (6th Cir.1989). Scientific test results, however, need not be infallible to meet the standard for due process. Dowling v. United States, 493 U.S. 342, 352-53, 110 S.Ct. 668, 674, 107 L.Ed.2d 708, 720 (1990). As the Supreme Court has stated, the due process standard only bars admission of evidence that is “so extremely unfair that its admission violates ‘fundamental conceptions of justice.’ ” Id. at 353, 110 S.Ct. at 674 (quoting United States v. Lovasco, 431 U.S. 783, 97 S.Ct. 2044, 52 L.Ed.2d 752 (1977) (citations omitted)). The Supreme Court has construed this test narrowly, id. at 352, 110 S.Ct. at 674, as have the Maryland courts. Crawford v. State, 285 Md. 431, 404 A.2d 244 (1979). For evidence to violate this standard because of its unreliability, the acts complained of must “be of such quality as necessarily prevent a fair trial.” Id. at 452, 404 A.2d at 255 (quoting Lisenba v. California, 314 U.S. 219, 62 S.Ct. 280, 86 L.Ed. 166 (1941)).

“Fundamental fairness” is an inherently malleable concept and, thus, does not lend itself to formulation of a bright-line rule. The issue of “fundamental fairness” must instead be assessed on the facts of each case. While there is neither a blanket test for fundamental fairness nor well-defined factors to guide trial courts, the Supreme Court has provided some broad considerations to weigh in assessing “fundamental fairness.” For example, in Dowling v. United States, 493 U.S. 342, 110 S.Ct. 668, 107 L.Ed.2d 708 (1990), the Petitioner was tried for an armed bank robbery in which the perpetrator had been wearing a ski mask and carrying a small pistol. A witness, Veronica Henry, testified that two weeks after the incident, the Petitioner had attempted to rob her, while wearing a ski mask and carrying a small pistol. Id. at 344-45, 110 S.Ct. at 670. Although Dowling had been acquitted of the robbery of Henry, the Court held that Henry’s testimony was not so “fundamentally unfair” that it violated Dowling’s due process rights. Id. at 353, 110 S.Ct. at 674. In rejecting Petitioner’s due process challenge, the Court emphasized the jury’s ability to weigh the evidence, and the Petitioner’s opportunity to challenge the testimony. Id.

Maryland courts have also considered the question of when evidence should be deemed so unreliable that it violates due process. The Court of Special Appeals has considered the issue, for example, in the context of paternity testing in Kammer v. Young, 73 Md.App. 565, 576-77, 535 A.2d 936, 941-42 (1988), cert. denied, 488 U.S. 919, 109 S.Ct. 298, 102 L.Ed.2d 318 (1988). In Kammer, the court considered a due process challenge to the reliability of calculations of the probability of the defendant’s paternity. By statute, paternity testing must exclude 97.3% of possible biological fathers to be admitted. Md.Code (1984, 1991 Repl.Vol., 1995 Cum.Supp.) § 5—1029(e)(1)(ii) of the Family Law Article. The defendant in Kammer argued that the conditional probability formula that was used to calculate the probability of his paternity was so unreliable that his due process rights were violated. Id. at 574, 535 A.2d at 940-01. The intermediate appellate court held, however, that his rights were not violated, explaining that:

Appellant was free to, and in fact did, put on non-genetic evidence which not only disputed generally his paternity but, in effect, was an attack upon the use of the ... [conditional] probability figure. This allowed him an opportunity to counterbalance appellee’s introduction of the blood test results and the prior probability on which they were based and served to protect his due process rights.

Id. at 577, 535 A.2d at 942. Cf. Wilson v. State, 70 Md.App. 527, 534, 521 A.2d 1257, 1262 (1987).

This Court has not previously considered the due process implications of the reliability of scientific evidence. In other contexts, however, we have considered whether evidence was sufficiently reliable to satisfy due process. For example, in Department v. Bo Peep, 317 Md. 573, 565 A.2d 1015 (1989), cert. denied, 494 U.S. 1067, 110 S.Ct. 1784, 108 L.Ed.2d 786 (1990), the Department of Human Resources held a hearing regarding revocation of Bo Peep Nursery’s license based on allegations of child abuse. In the hearing, the agency’s case against the nursery was based entirely on hearsay evidence. We held, however, that the exclusive use of hearsay, in the context of an administrative hearing, did not violate the nursery’s due process rights. Id. at 598-602, 565 A.2d at 1027-29. In so holding, we pointed to Bo Peep’s opportunity to cross-examine adult witnesses who spoke with the children. Id. at 601, 565 A.2d at 1028-29. We also contrasted this situation with cases where due process was violated because the defendant was completely unaware of the evidence his opponent intended to present, and thus had “no opportunity for cross-examination or rebuttal.” Id. at 598-99, 565 A.2d at 1027 (quoting Rogers v. Radio Shack, 271 Md. 126, 129, 314 A.2d 113, 115 (1974)).

From these cases, we distill the principle that the essence of the due process “fundamental fairness” inquiry is whether there was a balanced, fully explored presentation of the evidence. This balance in turn depends on the jury’s ability to weigh the evidence, and the defendant’s opportunity to challenge the evidence. See Dowling, 493 U.S. at 353, 110 S.Ct. at 674.

Applying these principles to the present case, we first consider the Petitioner’s generalized challenges to the DNA evidence, ie., the use of the product rule and the magnitude of the laboratory error rate compared to the odds of random matching. As we noted above, the most recent scientific data confirms that both the product rule and ceiling principle methods are viable, and therefore both are sufficiently reliable to satisfy due process. See supra Section V; see also VNTR Population Data Study, supra; Budowle et al., supra. We therefore conclude that because the jury was presented with both the product rule and ceiling principle calculations, with full explanations of both methods, it had the opportunity to weigh the contested evidence. In addition, although the Petitioner did not call independent experts at trial, we conclude that he had ample opportunity to challenge the product rule calculations.

We next turn to Petitioner’s argument that the magnitude of the laboratory error as compared to the odds of random DNA matching renders the odds of random matching meaningless. Essentially, Petitioner argues that the DNA testing procedure is inaccurate due to lab error in 0.7% of cases.³⁷ This means that in seven cases out of one thousand, an erroneous match may be found. Given this, he argues, it is improper to say that there is only one chance in 800,000 that the DNA match was “random” because there is at least a 0.7% chance of erroneous matching due to laboratory error. As we indicated above, however, the jury was fully informed of the laboratory error rate and the Petitioner had a full opportunity to address this on cross-examination. Therefore, there was no due process violation.

Finally, considering the Petitioner’s specific challenges to the DNA testing procedures used in his case, we first observe that although Petitioner characterizes his objection to the use of “match windows” as a case-specific challenge, it is really a general challenge to DNA testing. The use of match windows is an inherent component of the process of DNA testing. Regarding Petitioner’s other case-specific challenges, we conclude that these issues were fully presented to the jury at trial, and the jury was able to factor them into its assessment of the DNA evidence. The alleged technical defects related to the DNA testing were fully explained to the jury by the experts. We can not say that the data was so unreliable on its face that petitioner’s due process rights were violated. Under the circumstances presented herein, we find that these technical issues go to the weight of the DNA evidence, not its admissibility.

Petitioner had the opportunity to challenge the DNA evidence and to raise both components of potential error on cross-examination, and he did so. In discovery, the State provided all the background information the proponent of DNA evidence is required to provide under § 10—915(b)(1) & (b)(2), facilitating thorough cross-examination. Cf. Jackson v. State, 92 Md.App. 304, 322-23, 608 A.2d 782, 791 (1992) (petitioner raised a due process objection at trial, but dropped his constitutional arguments on appeal), cert. denied, 328 Md. 238, 614 A.2d 84 (1992). Therefore, his due process rights were not violated by presenting this thoroughly explored evidence to the jury.

JUDGMENT OF THE COURT OF SPECIAL APPEALS AFFIRMED. COSTS TO BE PAID BY THE PETITIONER.

1 . Unless otherwise specified, all statutory cites herein are to Maryland Code (1974, 1995 Repl.Vol., 1995 Cum.Supp.) § 10-915 of the Courts and Judicial Proceedings Article.

2 . In the trial at issue in this appeal, Armstead was tried on twelve of the twenty-five charges. Ten of the remaining charges were severed, two were dismissed, and one was not prosecuted.

3 . For purposes of the DNA hearing, this case was consolidated with the unrelated case of defendant John Daniel Kelly.

4 . Reed adopted the "general acceptance” standard originally set forth in Frye v. United States, 293 F. 1013, 1014 (D.C.Cir.1923), overruled by Daubert v. Merrell Dow, 509 U.S. 579, 588, 113 S.Ct. 2786, 2794, 125 L.Ed.2d 469, 480 (1993); hereinafter, we shall refer to this as the Frye-Reed standard.

5 . For a more detailed discussion of the mechanics and history of DNA testing in forensics, see generally H. Lee et al., DNA Typing in Forensic Science, 15 Am.J.Forensic Med. & Pathology 269 (1994); Committee on DNA Technology in Forensic Science, National Research Council, DNA Technology in Forensic Science (1992) (Prepublication Manuscript) [hereinafter NRC Report]; Office of Technology Assessment, Genetic Witness: Forensic Use of DNA Tests (1990). See also United States v. Yee, 134 F.R.D. 161, 169-73 (N.D.Ohio 1991), aff'd sub nom United States v. Bonds, 12 F.3d 540 (6th Cir.1993); State v. Vandebogart, 139 N.H. 145, 652 A.2d 671, 675-77 (1994); Com. v. Curnin, 409 Mass. 218, 565 N.E.2d 440, 445-48 (1991) (Appendix); People v. Castro, 144 Misc.2d 956, 545 N.Y.S.2d 985, 988-95 (Bronx County Ct.1989).

6 . Nearly all courts that have considered the admissibility of DNA evidence have found the general technique of DNA profiling reliable. See Developments in the Law: Confronting the New Challenges of Scientific Evidence, 108 Harv.L.Rev. 1481, 1558 (1994). Some courts, however, have rejected DNA evidence because the population genetics component was held to be unreliable, see, e.g., Com. v. Lanigan, 413 Mass. 154, 596 N.E.2d 311, 314 (1992) (product rule method of calculating odds of a random DNA match not generally accepted, and evidence of a DNA match inadmissible without supporting statistics) (Lanigan I), aff'd in part and rev’d in part, 419 Mass. 15, 641 N.E.2d 1342 (1994) (ceiling principle method of calculating odds of a random DNA match now generally accepted and, therefore, DNA evidence admissible) (Lanigan II); State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 494 (1992) (product rule method of calculating odds of a random DNA match not generally accepted, and evidence of a DNA match inadmissible without supporting statistics), modified on reh’g, 139 N.H. 145, 652 A.2d 671 (1994) (ceiling principle method of calculating odds of a random DNA match now generally accepted and, therefore, DNA evidence admissible if supporting statistics offered that were calculated using ceiling principle), or because of errors in applying the profiling methods in a particular case. People v. Castro, 144 Misc.2d 956, 545 N.Y.S.2d 985 (Bronx County Ct.1989).

7 . Even before it was used in forensics, DNA profiling had been used for a number of years for therapeutic purposes such as the diagnosis of hereditary medical diseases. Although there are significant distinctions between use of DNA analysis for forensic purposes versus therapeutic purposes, the longer experience with therapeutic DNA analysis has provided an opportunity to develop and refine the technique to its current level of consistency and reliability. See NRC Report, supra, at 2-2.

8 . The only situation where two people should be found to have identical DNA is in the case of identical twins, NRC Report, supra, at S-2, although siblings or other close relatives will also have substantial similarities in their DNA. See B. Bockel et al., Likelihoods of Multilocus DNA Fingerprints in Extended Families, 51 Am.J.Hum.Genetics 554, 559 (1992).

9 . Although most cases to date have involved the RFLP method of DNA analysis, a newer method known as polymerase chain reaction (PCR) has also been developed. NRC Report, supra, at 1-8. This method is particularly useful in analyzing DNA where there is a very small evidence sample to be tested, because it utilizes the same method by which cells replicate DNA to “amplify” the quantity of DNA. Id. Alex Jeffreys, a pioneering DNA researcher, has proposed a digital technique for applying PCR that would eliminate some of the frequently challenged aspects of RFLP testing. Id. at 1-10. The PCR method, however, is subject to other types of technical error. Id. at 2-14 to 2-24.

Courts have already begun to evaluate the PCR method, and some have already held that the technique is generally accepted as reliable. See, e.g., State v. Gentry, 125 Wash.2d 570, 888 P.2d 1105, 1117 (1995); People v. Lee 212 Mich.App. 228, 537 N.W.2d 233 (1995).

10 . Sections 10-915b(1) and b(2) of the statute provide that DNA profile evidence is admissible if the proponent:

(1) Notifies in writing the other party or parties by mail at least 45 days before any criminal proceeding; and

(2) Provides, if requested in writing, the other party or parties at least 30 days before any criminal proceeding with:

(i) Duplicates of the actual autoradiographs generated;

(ii) The laboratory protocols and procedures;

(iii) The identification of each probe utilized;

(iv) A statement describing the methodology of measuring fragment size and match criteria; and

(v) A statement setting forth the allele frequency and genotype data for the appropriate data base utilized.

11 . By comparison, Virginia’s statute governing the admissibility of DNA evidence provides that "DNA ... testing shall be deemed to be a reliable scientific technique and the evidence of a DNA profile comparison may be admitted to prove or disprove the identity of any person.” Va.Code Ann. § 19.2-270.5 (Michie 1995) (emphasis added). The conditional language of the Virginia statute has been interpreted to permit trial judges to continue to exercise their discretion to weigh the probative value of the DNA evidence against its prejudicial effect. Satcher v. Com., 244 Va. 220, 421 S.E.2d 821, 835 (1992), cert. denied, 507 U.S. 933, 113 S.Ct. 1319, 122 L.Ed.2d 705 (1993). See infra Section IV.B.2.

12 . Maryland is one of twelve states to enact legislation establishing the admissibility of DNA evidence in criminal cases. In addition to Maryland, the states that have enacted such legislation are: Alaska, 1995 Alaska Sess.Laws 7, § 2 (codified at Alaska Stat. § 12.45.035 (1995)); Connecticut (Conn.Gen.Stat. § 54-86k (1995)), Delaware (Del.Code Ann. tit. 11, § 3515 (Michie Supp.1994)), Indiana (Ind.Code Ann. § 35-37-4-13 (Burns 1994)), Louisiana (La.Rev.Stat.Ann. § 15:441.1 (West Supp.1995)), Minnesota (Minn.Stat.Ann. §§ 634.25-634.26 (West Supp. 1995)), North Dakota (N.D.Cent.Code § 31-13-02 (Michie Supp.1995)), Nevada (Nev.Rev.Stat.Ann. § 56.020 (Michie 1986 & Supp. 1993)), Tennessee (Tenn.Code Ann. § 24-7-117 (Michie Supp.1995)), Wisconsin (Wis.Stat.Ann. § 972.11 (West Supp.1994)), and Virginia (Va.Code Ann. § 19.2-270.5 (Michie 1995)).

Four of the other eleven states—Alaska, Minnesota, Nevada, and Tennessee—have enacted statutes that explicitly state that probability estimates are also admissible. These provisions, however, take different forms. Tennessee's statute expressly states that "statistical population frequency evidence ... is admissible in evidence to demonstrate the fraction of the population that would have the same combination of genetic markers as was found in a specific biological specimen.” Tenn.Code Ann. § 24-7-117(c) (Michie Supp.1995). Nevada’s statute provides more generally that "[t]he opinion of any expert concerning results of blood tests may be weighted in accordance with evidence, if available, of the statistical probability of the alleged blood relationship.” Nev.Rev.Stat.Ann. § 56.020 (Michie 1986 & Supp.1993). Alaska’s statute defines "DNA profile” to include "statistical population frequency comparisons of the patterned chemical structures.” 1995 Alaska Sess. Laws 7, § 2 (codified at Alaska Stat. § 12.45.035(b)(2)(B) (1995)).

Minnesota is unique because although its statute explicitly provides for the admission of statistics in support of DNA evidence, its courts have not uniformly admitted the statistics. The Minnesota statute provides that: “statistical population frequency evidence ... is admissible to demonstrate the fraction of the population that would have the same combination of genetic markers as was found in a specific biologic specimen.” Minn.St.Ann. § 634.26 (West Supp.1995). In Minnesota, a line of cases culminating with State v. Kim, 398 N.W.2d 544 (Minn. 1987) had imposed limitations on the use of statistics because of the potentially “exaggerated impact on the trier of fact.’’ Id. at 548. The Minnesota Supreme Court in State v. Nielsen, 467 N.W.2d 615 (Minn.1991), questioned the legislature's authority to create an exception to the Kim doctrine for DNA evidence. Id. at 620. Relying on Nielsen, an intermediate appellate court again questioned the legislature’s authority to render the statistics admissible in State v. Alt, 504 N.W.2d 38, 41 n. 2 (Minn.App.1993), aff’d as modified, 505 N.W.2d 72 (Minn.1993). More recently, in State v. Bloom, 516 N.W.2d 159 (Minn. 1994), the Minnesota Supreme Court modified its position to permit expert witnesses to provide statistical evidence in DNA cases, but only under the ceiling principle approach proposed in the NRC report. Id. at 167.

13 . While the bill as originally proposed would have given parties the right to depose any witness testifying about the DNA evidence, this provision was eliminated prior to enactment. 1991 Maryland Laws ch. 631.

14 . One of the amendments made to the DNA legislation after it was proposed and before its passage was to delete the words "unique” and "uniquely” throughout the bill whenever they were used to refer to an individual’s DNA pattern. 1989 Maryland Laws ch. 430, § 1, at 2893 (1989) (codified as amended at § 10-915). This alteration demonstrates that the Legislature was aware of the possibility that an individual's DNA pattern was not unique, and thus recognized the potential for random matching. Nevertheless, the Legislature unequivocally deemed DNA profiling reliable.

15 . The NRC Report estimated that states would spend millions of dollars on forensic DNA evidence, including expenditures to provide experts to testify for the State, to fund testing and experts for indigent defendants, and to maintain and upgrade DNA databases storing the profiles of convicted felons. NRC Report, supra, at 6-18 to 6-19.

16 . Paternity testing cases also serve to clarify the proper role of the trial court in determining relevance. For example, in Haines v. Shanholtz, 57 Md.App. 92, 468 A.2d 1365 (1984), cert. denied, 300 Md. 90, 475 A.2d 1201 (1984), as here, the Legislature enacted a statute governing admissibility of scientific evidence. Md.Code (1984, 1991 Repl.Vol., 1995 Cum.Supp.) § 5-1029(e)(ii) of the Family Law Article. In Haines, however, the evidence at issue was blood testing. The statute provided that “test results may be received in evidence in cases where ... the probability of the alleged father’s paternity is at least 97.3 percent.” Id. at 96, 468 A.2d at 1366-67. Under this statute, the Court of Special Appeals held that:

[T]he trial court should not have engaged in a determination of scientific acceptance of genetic testing for establishing paternity ... The legislative intent is clear—genetic testing is valid, subject to cross-examination of those responsible for performing and evaluating the tests. The trial judge has the discretion of admitting testimony that may have a bearing on the weight of the testimony offered ... but not, by reason of the statute, as to the admissibility vel non of genetic testing to establish paternity.

Id. at 97-98, 468 A.2d at 1367 (emphasis in original). See also Kammer v. Young, 73 Md.App. 565, 535 A.2d 936 (1988), cert. denied, 488 U.S. 919, 109 S.Ct. 298, 102 L.Ed.2d 318 (1988).

17 . The court in Castro articulated a three-part test for admissibility of DNA evidence:

Prong I. Is there a theory, which is generally accepted in the scientific community, which supports the conclusion that DNA forensic testing can produce reliable results?

Prong II. Are there techniques or experiments that currently exist that are capable of producing reliable results in DNA identification and which are generally accepted in the scientific community?

Prong III. Did the testing laboratory perform the accepted scientific techniques in analyzing the forensic samples in this particular case?

545 N.Y.S.2d at 987.

18 . This approach is supported by the NRC Report, which stated that:

The validity of [the] assumption ... that the analytical work done for a particular trial comports with proper procedure ... can be resolved only case by case and is always open to question, even if the general reliability of DNA typing is fully accepted in the scientific community. The DNA evidence should not be admissible if the proper procedures were not followed. Moreover, even if a court finds DNA evidence admissible because proper procedures were not followed, the probative force of the evidence will depend on the quality of the laboratory work.

NRC Report, supra, at 6-4.

In determining whether an alleged error in DNA testing constitutes the type of error that warrants exclusion of DNA evidence, trial courts must distinguish mere measurement error, which is inherent in any scientific procedure, from deviations from accepted testing procedures. For example, contamination or degradation of the DNA sample constitutes the type of error that warrants exclusion.

19 . The Petitioner also raised another challenge, i.e., that the differences in the results obtained by the FBI testing laboratory and Cellmark Diagnostics rendered the test results in his case so unreliable that they lacked probative value. The specific difference was that the FBI autoradiograph showed a double band at one locus, while Cellmark's autoradiograph only showed one band at that locus. The presence of two bands at one locus generally indicates that a person is heterozygous, which means they have two different alleles or forms of a particular gene. G. Beadle & M. Beadle, the Language of Life 54-66 (1966). For example, if a person has one gene for blue eyes and another gene for brown eyes, he or she is heterozygous for the eye color gene. By comparison, the presence of only one band may indicate that a person is homozygous, or that they have two copies of the same allele for a particular gene. Id. Absent a rare genetic mutation, the same person can not be both homozygous and heterozygous for a given gene, since this is determined at conception, when each parent contributes one allele of each gene. Id. at 65.

As the experts explained in the pre-trial hearing, however, this discrepancy can be accounted for by differences in the testing conditions used by the two laboratories. The FBI expert, Dr. Dwight Adams, and the Cellmark expert, Dr. Charlotte Word, attributed the difference to the likelihood that low-molecular weight bands may run off the bottom of the gel, causing only one band to appear at a locus where two bands would normally be found. Dr. Word noted that Cellmark’s procedures typically yielded larger fragments, while the FBI’s system typically yielded smaller fragments. As a result, the Cellmark method would be more likely to show very large bands, which would often be lost using the FBI methods, while the FBI system would be more likely to show very small bands.

Although Armstead frames this argument as one that is specific to the DNA evidence in his case, we believe it is really a general attack on the inherent variability of DNA analysis. All DNA testing will vary slightly depending on the type of probe, the restriction enzyme, the gel, the amount of time the experiment "runs,” etc. General attacks on these procedures are precluded under § 10-915.

20 . In reality, two alleles would be identified at each locus tested. For example, at locus one, a person could be found to have one of three "genotypes” or combinations of alleles: A1/A1; A1/B1, or B1/B1. NRC Report, supra, at 1-5 (Fig. 1-3). For simplicity, however, our hypothetical probability calculations use only one allele at each locus.

21 . A third question has been raised somewhat less often than the two discussed above regarding DNA evidence. This issue relates to the mechanics of how the visual pattern produced by RFLP analysis is translated into allele assignments. As described above, the variable areas of DNA tested in DNA profiling, VNTR’s, may contain between 20 and 100 repeats of the same nucleotide sequence. One profile may have a 29-repeat segment, while another may have a 35-repeat segment. DNA profiling, however, is not sufficiently precise to identify small differences in the number of repeats; therefore, a process known as "binning” is used to sort the differing lengths of DNA fragments into categories. Each category is interpreted to represent a different allele, although in reality, each may contain several alleles: for example, one "bin” may include 26-repeat, 27-repeat, and 28-repeat segments because their lengths are not sufficiently different to distinguish them. See, e.g., L. Mueller, Population Genetics of Hypervariable Human DNA, in Forensic DNA Technology 56 (1992). Several conservative assumptions have been incorporated in the binning process, however, which have resolved most theoretical objections, although concerns about binning have not completely disappeared.

22 . For a thorough discussion of independent and non-independent events, see D. Stirzaker, Elementary Probability 22-30 (1994). Stirzaker provides a useful genetic example at pages 29-30. Id.

23 . The terms “linkage equilibrium” and "gametic phase balance” are related to the concept of Hardy-Weinberg equilibrium, although they are not interchangeable. All reference a theoretical state of random mating across racial and ethnic lines. Lewontin & Hartl, supra, at 1746-47.

24 . The ceiling principle “assumes the worst” about the amount of variation in allele frequencies across subpopulations. Under the product rule, the odds of a random match at one allele are determined by comparison with the frequency of occurrence of that allele in the population. Therefore, if a particular allele that shows up in the suspect’s sample occurs in 1% of the population, there is a 1% chance of a random match.

Suppose, instead, that the allele in question occurs in a low percentage of Caucasians, but in a much higher percentage of Hispanics. The ceiling principle proponents at one point suggested that, in case such variability existed (although data did not then exist to confirm or refute this), we should assume a much higher maximum frequency than is likely, so that the possibility of error due to a random match is minimized. NRC Report, supra, at 3-10 to 3-12. The NRC report recommended using an "interim” population allele frequency of 10%, to be replaced by a figure of 5% once preliminary data was collected. Id. at 3-21. These rates were considerably higher than the likely maximum allele frequency for any subgroup, based even on then-existing data. See, e.g., B. Devlin et al., Statistical Evaluation of DNA Fingerprinting: A Critique of the NRC’s Report, 259 Science 748, 749 (1993).

Other than incorporating these maximum allele frequencies, the product rule would still be applied as discussed above. See, e.g., NRC Report, supra, at S-l 1 to S-12.

25 . Emphasizing the convergence of scientific opinion regarding population genetics statistics, Lander and Budowle note in their article that:

As co-authors, we can address these questions in an even-handed manner. B.B. [Bruce Budowle] was one of the principal architects of the FBI’s DNA typing programme, whereas E.S.L. [Eric S. Lander] was an early and vigorous critic of the lack of scientific standards and served on the NRC committee. In a world of soundbites, we are often pegged as, respectively, a "proponent” and an "opponent” of DNA typing. Such labels greatly oversimplify matters, but it is fair to say that we represent the range of scientific debate.

Lander & Budowle, supra, at 735.

26 . While Hartl and Lewontin condnue to advocate conservative treatment of DNA evidence, their attack on population genetics statistics has shifted; Lewontin, for example, recently asserted that "juries are no more capable of understanding probability statements than they are of interpreting any other piece of highly technical information.” R. Lewontin, Letter, Nature 398 (1994); see also D. Hartl, Letter, 372 Nature 398-99 (1994). This position has been subject to considerable criticism by other scientists. For example, in a recent response to Lewontin's comment, the author stated "[t]he continued existence of a Flat Earth Society and the increasing popularity of Creationism demonstrate that it is never possible to convince every individual of the validity of a scientific theory,” and noted how "a tiny, vocal minority with access to media outlets can attempt to sway public opinion against generally accepted medical and scientific opinions.” C. Strom, Letter, 373 Nature 98-99 (1995).

27 . The dissent apparently dismisses the importance of the FBI study in dispelling the controversy regarding the product rule versus the ceiling principle. See dissenting op. at 105-106 & n. 7. Contrary to the dissent's view, however, courts that have addressed the admissibility of population genetics subsequent to the FBI’s VNTR study have recognized that current scientific data does not support the need to apply either the ceiling principle or the modified ceiling principle. See, e.g., Smith, 49 Cal.Rptr.2d at 613-14; Marlow, 41 Cal.Rptr.2d at 32-33.

28 . This approach, however, also presents problems, because

[t]o use the specific ethnic background of the suspect (which may be impossible to define) would presuppose that he or she be the true perpetrator. However, if the true perpetrator were known a priori, there would be no need for statistical estimates. Furthermore, if a particular subgroup was chosen as the reference database, for the majority of cases this would insinuate that a member of one subgroup is a more likely source of the crime scene evidence.

Budowle et al., supra, at 320. The authors therefore conclude that "[sjince the ethnicity of those people who are potential perpetrators rarely, if ever, is known, statistical estimates must be based on some sort of general population database.” Id.

29 . In the present case, population genetics statistics were calculated using a Caucasian database, a Hispanic database, and an African American database. The odds of one in 800,000 and one in 480 million referred to in this appeal were calculated using the African American database. The database consisted of approximately 250 persons from the Detroit metropolitan area.

30 . Lempert argues that while population substructuring is no longer a significant concern in most cases, there are still caveats regarding the use of DNA evidence. R. Lempert, The Suspect Population and DNA Identification, 34 Jurimetrics J. 1 (1993). For example, he suggests that the odds of random matches between the suspect DNA and DNA taken from the crime scene will be underestimated where the suspect population includes one or more of the suspect’s close relatives. Id. at 6. The problem is greatest in small, isolated populations with atypically constrained mating, e.g., in the Pennsylvania Amish community, certain Native American tribes, etc. Id. at 2-3.

This argument arose in United States v. Two Bulls, 918 F.2d 56 (8th Cir.1990), vacated and dismissed as moot, 925 F.2d 1127 (8th Cir.1991) (en banc). The defendant in Two Bulls was a Native American, and the relevant suspect population included a large number of people with the same tribal background as the defendant. See Lempert, supra, at 5, n. 9.

31 . Among the courts that have found statistical evidence inadmissible, there have been at least three different approaches to the use of DNA "match” testimony:

(1) Disallow use of the DNA match as well, because it is meaningless without contextual statistics;

(2) "Uncouple” the match evidence from the statistical evidence, permitting testimony as to the match; or

(3) Permit testimony regarding the match, and allow expert testimony regarding the frequency of occurrence of each allele in the general population, but disallow testimony giving an overall probability of match (thus, avoiding the product rule issue, because the probabilities are not combined).

See State v. Bible, 175 Ariz. 549, 858 P.2d 1152, 1190 (1993), cert. denied, - U.S. -, 114 S.Ct. 1578, 128 L.Ed.2d 221 (1994).

32 . To the extent that Jackson v. State, 92 Md.App. 304, 608 A.2d 782 (1992), cert. denied, 328 Md. 238, 614 A.2d 84 (1992), is inconsistent with this holding, it is hereby disapproved. In Jackson, the defendant argued that expert testimony that his DNA matched DNA from the crime scene should not have been admitted because no population genetics statistics were provided to put the “match” testimony in context, asserting that “without proper evidence regarding the probability of a match, evidence that a match was declared has no relevance. Without probability calculations the fact that there was a match does not tend to make it more or less likely that [a]ppellant was the assailant.” 92 Md.App. at 324, 608 A.2d at 791. While the court found that this argument had been waived, it noted that: “In any event, we have explained that DNA testing has been legislatively determined to be reliable and is generally admissible in Maryland. The expert witness testified that she used standard procedures and standard equipment in conducting the testing and comparisons. There was simply no need for the State to offer additional evidence, such as probability calculations, to establish that the testing procedures employed were reliable.” Id. at 324, 608 A.2d at 792.

33 . In Keirsey, the Court of Special Appeals held that the Frye-Reed test was inapplicable to the methods of calculating population genetics statistics, stating that "the Frye-Reed test ... is applicable only when an essential component of the expert's opinion is a scientific test result ‘controlled by inexorable, physical laws.'” 106 Md.App. at 575, 665 A.2d at 712 (quoting State v. Allewalt, 308 Md. 89, 98, 517 A.2d 741, 745 (1986)). We believe this description of the scope of the Frye-Reed test is too limited, excluding some mathematical techniques that should be subjected to reliability analysis.

The Frye-Reed test often will not apply to statistical calculations because the choice between alternative statistical techniques, although subjective, is often merely a choice between equally valid methods of describing the same underlying scientific data. Statistics are inherently flexible, and thus there are usually multiple correct statistics that can be used to describe the same set of data. Statisticians routinely make choices in presenting data; for example, they may choose to present either the mean, the median, or the mode to describe the "center" of a data set. This type of format choice is not subject to Frye-Reed analysis.

There are, however, instances, as in this case, where the proper choice of statistical techniques is dependent on an underlying scientific phenomenon or principle. For example, suppose that a new species of flower is discovered. When it is discovered, a white-flowered variety and a red-flowered variety are observed. It would be incorrect to calculate the probability of a new plant having white flowers based on a normal distribution, because this would depend on whether flower colors varied along a continuum from white to pink to red, or whether there were only discrete possibilities for the flower color, i.e., white or red. See R. Freund & W. Wilson, Statistical Methods 65-66, 70-76. Under this scenario, the correct choice of probability calculations would depend on the underlying genetics of the plant.

We believe the choice between population genetics approaches—i.e., the product rule versus the ceiling principle—is similarly dependent on an underlying scientific hypothesis, because if the scientific data showed significant variation in allele frequencies across subpopulations, this would imply that use of the product rule was incorrect. Other courts have also adopted this view. For example, in People v. Soto, 39 Cal.App.4th 757, 35 Cal.Rptr.2d 846 (1994), cert. granted, 39 Cal.Rptr.2d 406, 890 P.2d 1115 (1995), a California court determined that the Frye standard applied to the population genetics formulae used to analyze DNA evidence. Id. 35 Cal.Rptr.2d at 858. Because the scientific data has shown the variation in allele frequencies to be insignificant, however, the Soto court ultimately concluded that both the product rule and the ceiling principle passed the “general acceptance” test. Id. See also E. Immwinkelried, The Methods of Attacking Scientific Evidence § 6.6(D), at 228-29 (1982).

Although we therefore disagree with the Court of Special Appeals' circumscribed view of the Frye-Reed test, we need not carry our analysis as far as in Soto, because the Maryland General Assembly has determined that statistical evidence of the odds of a random DNA match is admissible.

34 . Taken to the extreme, this approach would require the Legislature to specify not only the type of statistical formula to be applied, but also the type of probe to be used, the number of loci to be tested, the minimum database size, and a multitude of other details which impact on the accuracy and reliability of the results.

35 . While the court in Bible acknowledged the evolutionary nature of scientific development, observing that "neither logic nor authority supports confining ourselves to a snapshot, rather than viewing the motion picture, of technological advancement,” it used this principle to support a different conclusion than the one we reach here. 858 P.2d at 1189 n. 33. The Bible court concluded that on appeal, review of the trial court’s Frye-Reed analysis should not be limited to the level of acceptance of DNA testing at the time of trial, but rather should also consider subsequent scientific developments. Id. The Bible court, however, was not determining the scope of admissibility under a statute, but rather considered the admissibility of DNA evidence in a jurisdiction without legislation governing admissibility.

36 . The FBI will declare a match between two fragments of DNA if the size of one fragment is within 2.5% of the size of another. To illustrate, if fragment A is 10,000 nucleotides long, while fragment B is 10,250 nucleotides long, they will still be declared a “match.” Cellmark uses a match window that varies with the size of the DNA fragment.

37 . The laboratory error rate of 0.7% for Cellmarlc was based on two errors identified in proficiency tests conducted in 1988. Since that time, the laboratory protocol has been revised to address these problems.