The defendant, Daniel Vandebogart, appeals from his conviction for first degree murder, RSA 630:1-a, based on a jury verdict in the Superior Court (Mohl, J.). Upon motion by the State, we bifurcated the defendant’s appeal in order to expedite our consideration of issues he raises relating solely to the admissibility of forensic DNA profiling. The defendant complains that the trial judge misapplied the legal standard for admitting novel scientific evidence under Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), and State v. Coolidge, 109 N.H. 403, 260 A.2d 547 (1969), rev’d on other grounds, 403 U.S. 443 (1971). We reverse and remand.
On September 12, 1989, Kimberly Goss was raped and murdered. In October, the New Hampshire State Police asked the FBI to perform forensic DNA analysis. Subsequently, the FBI’s DNA laboratory received a package of forensic samples which included known blood samples from Kimberly Goss and the defendant together with two vaginal swabs taken from Kimberly Goss at her autopsy. On February 27, 1990, the FBI’s DNA laboratory reported to the New Hampshire State Police that the genetic profile of the defendant’s blood sample matched the genetic profile of semen found on the vaginal swabs at three genetic locations. The FBI further reported that the probability that an unrelated individual selected at random from the Caucasian population would have a genetic profile matching the defendant’s at those three locations was 1 in 50,000.
Prior to trial, the defendant filed a motion in limine seeking to exclude the DNA evidence of a match and the probability calculation. *367In response to the defendant’s motion, the trial court held a pretrial Frye hearing which lasted ten days. At the hearing, the court heard expert testimony from five witnesses for the prosecution and three for the defendant. The State’s witnesses and their backgrounds were as follows: (1) Dr. Dwight Adams, a biologist, a member of the American Academy of Forensic Sciences, and the FBI special agent responsible for the DNA analysis in this case; (2) Dr. Steven Daiger, Professor of Medical Genetics at the University of Texas Health Science Center; (3) Dr. David Goldman, Chief of the Genetics Studies Section of the National Institutes of Alcohol Abuse and Alcoholism, National Institute of Health; (4) Dr. Michael Conneally, Distinguished Professor of Medical Genetics and Neurology at Indiana University Medical Center; and, as a rebuttal witness, (5) Dr. Bruce Budowle, a molecular biologist and human population geneticist in charge of the FBI’s DNA research program. The following witnesses testified for the defendant: (1) Dr. William Shields, Professor at State University of Environmental Science Technology, Syracuse, New York; (2) Dr. Joseph Nadeau, a scientist working at Jackson Laboratory and recipient of a Ph.D. in population genetics from Boston University; and (3) Dr. Everett Mendelsohn, a Professor of History of Science at Harvard University with a Ph.D. in the History of Science.
Following the hearing, the court issued an order denying the defendant’s motion. After a trial in which the evidence derived from the DNA testing was admitted, the jury convicted the defendant of first degree murder, and the court sentenced him to life in prison without parole.
In this portion of the defendant’s bifurcated appeal, the sole issue for our consideration is whether the trial court properly applied the standard for admissibility of novel scientific evidence. The defendant argues here, as he did below, that the proper test for admissibility of novel scientific evidence requires a three-prong analysis under Frye and Coolidge, and that the trial court erred by not applying the second and third prongs. Specifically, he contends that the trial court only examined the general acceptance of the theory underlying DNA profiling, and that if it had properly applied the second prong, it would have found that the particular technology used by the FBI to perform the DNA profiling analysis was not generally accepted as reliable by the relevant scientific community. The State responds that the trial court properly applied the Frye standard and that in its order the court correctly found that both the theory and technology of DNA profiling were generally accepted.
*368 I. DNA Background
A basic understanding of the theories and procedures involved in DNA profiling is necessary to understand the legal issues surrounding its use as evidence in court. Therefore, before we discuss the issues surrounding the admissibility of novel scientific evidence, we shall first consider the general nature of the particular evidence the State sought to have admitted. We derive our scientific exposition of DNA and DNA profiling from testimony given at the Frye hearing and from a report entitled “DNA Technology in Forensic Science,” which the National Research Council published in April 1992. For more comprehensive descriptions of these topics, see United States v. Jakobetz, 747 F. Supp. 250, 250-54 (D. Vt. 1990), aff'd, 955 F.2d 786 (2d Cir. 1992), cert. denied, — U.S. —, 61 U.S.L.W. 3257 (Oct. 5, 1992); People v. Wesley, 533 N.Y.S.2d 643, 645-50 (County Ct. 1988); E. Imwinkelried, The Debate in the DNA Cases over the Foundation for the Admission of Scientific Evidence: The Importance of Human Error as a Cause of Forensic Misanalysis, 69 WASH. U.L.Q. 19 (1991); W. Thompson & S. Ford, DNA Typing: Acceptance and Weight of the New Genetic Identification Tests, 75 Va. L. Rev. 45 (1989).
A. DNA Theory
Deoxyribonucleic acid (DNA) is an organic substance found in the chromosomes contained in the nucleus of a cell. It provides the genetic blueprint that determines the physical structures and individual characteristics of every living organism—humans, animals, plants, and even bacteria. In humans, DNA exists in all cells that have a nucleus, including white blood cells, sperm, cells surrounding hair roots, and the cells in saliva. These are the cells most often discovered at crime scenes and are the most useful in forensic DNA analysis.
With exceptions not relevant here, DNA does not vary within an individual, i.e., the DNA contained in one cell in an individual will be identical to the DNA contained in every other cell of that individual. For forensic purposes, the important characteristic of DNA is that, excepting identical twins, no two persons have the same DNA structure.
The DNA molecule is shaped like a double helix which resembles a twisted ladder. Each component strand of the helix, similar to the rungs on a ladder, consists of a sequence of nucleotides. The nucleotides are sometimes referred to as bases. There are four types of nucleotides in the DNA molecule, and they are designated as adenine *369(A), guanine (G), eystonine (C), and thymine (T). The nucleotides bond in predictable patterns, A to T and C to G. A pair of complementary bases—A-T, T-A, C-G, or G-C—is designated as a base pair. The order in which these base pairs appear on the DNA ladder constitutes the genetic code for the cell. This code carries the necessary information to produce the many proteins which comprise the human body. A sequence of base pairs responsible for producing a particular protein is called a “gene.” A gene, the basic unit of heredity, consists of a sequence of between 1,000 and 2 million nucleotides. Scientists estimate that the human genome, the complete genetic makeup of a person, contains 50,000 to 100,000 genes and that in a human set of 23 chromosomes there are about 3 billion nucleotides.
Inheritable characteristics are controlled by pairs of genes, or alleles, which occupy the same sites, or loci, on paired chromosomes. One of each pair of alleles is inherited from the father, and one is from the mother. When the alleles that comprise a pair differ, the individual is said to be “heterozygous” for that allele. When the maternal and paternal alleles in a pair are the same, the individual is “homozygous.” A particular combination of alleles is referred to as a genotype.
DNA technology allows scientists to detect genetic variations. A characteristic that differs among individuals is termed a polymorphism. In DNA profiling, the terms polymorphism and variation are used interchangeably. Some regions of DNA contain repetitive strings of nucleotides that are highly polymorphic. These are called “variable number tandem repeats” (VNTRs). At VNTRs, the number of repetitions of a nucleotide sequence can vary among individuals. For this reason, VNTRs are commonly used as genetic markers to detect variations.
A variation of even one nucleotide in the sequence of DNA is detectable. Such a variation can be detected by applying a biological catalyst, called a “restriction enzyme,” to the DNA. The restriction enzyme will cut the DNA into fragments of different lengths depending on the cutting sites recognized by the enzyme. These fragments of varying lengths are called “restriction fragment length polymorphisms” (RFLPs). Differences among individuals can be detected by the differences in the lengths of restriction fragments. Because of its extensive variability, the VNTR class of RFLPs is the most useful in distinguishing among individuals.
B. DNA Profiling Techniques
DNA profiling can inculpate a criminal suspect by comparing the suspect’s genetic material with genetic material obtained from hu*370man tissue left at the crime scene. DNA profiling involves two distinct procedures. First, RFLP analysis determines if there is a “match.” A “match” does not mean that the suspect was definitely the source of the genetic material found at the crime scene, however, but simply that the suspect cannot be eliminated as the potential source. Even if there is a perfect match, there is a possibility that the two samples came from different people whose DNA patterns at the targeted loci are indistinguishable. Thus the second procedure, population frequency calculation, generates a ratio which accompanies a match in order to express the statistical likelihood that an unrelated individual chosen at random from a particular population could have the same DNA profile as the suspect.
DNA analysis is generally performed by first disassembling the DNA molecular ladder in one of several different ways. The FBI uses “RFLP analysis,” and follows a written protocol that requires certain procedures for quality control and verification. The operative steps of RFLP analysis are outlined below:
1. Extraction of DNA. The DNA is first extracted from the evidentiary sample by using chemical enzymes and then purified.
2. Restriction of digestion. The DNA is then cut with chemical scissors called “restriction endonucleases.” These endonucleases recognize certain base pairs and sever the DNA molecule at specifically targeted base pair sites to produce RFLPs.
3. Gel electrophoresis. The cut fragments of DNA molecules are next placed in an agarose gel which is later electrically polarized to sort the fragments by length. Because DNA is negatively charged, the fragments will migrate toward the positive end of the gel. The distance traveled will depend upon the length of the fragment, with the shorter fragments traveling further in the gel. Molecular weight standards, also called “size markers,” are placed in separate lanes to measure the distance that the fragments travel. For comparison, several different samples of DNA from known and unknown sources are run on the same gel, but in different tracks or lanes.
A Southern blotting or transfer. Because the agarose gel is very difficult to work with, the fragments are transferred to a more functional surface by a method called “Southern transfer.” A nylon membrane is placed over the gel, which is set upon a sponge saturated with sodium hydroxide solution. The solution carries the fragments from the gel onto the nylon membrane, and they become permanently fixed on the membrane, referred to as a “blot,” in the same pattern as in the gel. Also during this step, a denaturation process severs each double-stranded DNA fragment into two single strands.
*3715. Hybridization. Next, a single-locus genetic probe is used to locate a specific polymorphic region of the DNA on the blot. A genetic probe is a single-stranded segment of DNA designed to complement a single-stranded DNA base sequence that is associated with a particular locus on a chromosomal pair. The probe will bond with any single-stranded fragments containing that particular base sequence. The typical result is that the probe will bind to DNA fragments at one or two locations in each lane, depending on whether the individual is homozygous or heterozygous for that particular allele. The genetic probe is tagged with a radioactive marker, which attaches to the probe and emits radiation without altering the function of the probe. The marker is used to determine the probe’s position on the blot after it hybridizes with polymorphic segments.
6. Autoradiography. Autoradiography is the photographic process that reveals the position of the polymorphic DNA segments. After hybridization, the nylon membrane is placed between two pieces of X-ray film. The radioactive probes expose the film at their respective locations. Black bands appear on the processed film where the radioactive probes have bonded to the RFLPs, producing a DNA “print.” Typically, each probe will expose one or two bands for each DNA sample, which reflects the maternal or paternal contributions to the individual’s DNA profile. The position of each band indicates the location of a polymorphic segment on the blot. Location, in turn, indicates the length of the DNA fragment that contains the segment. Because the length of the DNA fragments varies among individuals, the position of their bands on a DNA print can differentiate individuals.
After the first probe has been applied and the autoradiography process is complete, the first probe is stripped from the membrane. The hybridization process is then repeated on the same membrane using a second probe. This process is designed to locate a different VNTR base sequence on another chromosomal pair. The FBI usually repeats the hybridization and autoradiography processes using four or five different probes sequentially on a single blot. Repeating the processes with different probes decreases the likelihood that a match between the defendant’s profile and the forensic profile is a random event. It is rare for two unrelated persons to have eight or ten matching alleles across four or five different VNTR loci.
7. Interpretation of autoradiographs. The final step is to determine if a match exists in the two lanes of the autoradiograph between the DNA sample taken from the suspect and the forensic sample taken from the crime scene or victim. The FBI uses a two-stage pro*372cedure for deciding whether a match exists. First, the FBI looks for a visual match. A visual match means that the forensic sample of DNA and the suspect’s DNA have the same number of bands in approximately the same locations on each autoradiograph. If no visual match exists, the FBI decides whether the non-match should be interpreted as inconclusive or as excluding the suspect. If a visual match is declared, the FBI uses a computer-assisted process to verify the existence of a match. Through a series of calculations, the computer will determine whether the difference in size of the fragments detected in the defendant’s sample and the forensic samples is within accepted limits. If the size of the suspect’s DNA fragments and the forensic samples are within plus or minus two and one-half percent of each other, then the visual match is confirmed. If the difference between the two exceeds the “matching criteria” of plus or minus two and one-half percent, then the autoradiograph is considered either inconclusive or as excluding the suspect. In this case, the FBI confirmed a visual match between the defendant’s DNA and that from the victim’s body because the degree of variation did not exceed plus or minus one percent.
Once the suspect’s DNA profile is declared to match the forensic sample, the FBI relies on statistical methods used in population genetics to calculate the likelihood of a random match. “Fixed bin analysis” is the FBI’s method for assigning to each band in a DNA profile a value or frequency that represents how often a particular allele may occur at a specific VNTR locus in a given population. To estimate population frequencies for particular alleles at targeted VNTR loci, the FBI has compiled data bases for Caucasian, Black, Asian, and Hispanic populations. The FBI’s Caucasian data base was derived from RFLP analyses of blood samples of approximately 225 FBI agent-trainees. The end result of the FBI’s fixed bin analysis of RFLPs from a forensic sample is a statistic which estimates the probability that the DNA profile of an individual chosen at random from a given population might match the DNA profile of the forensic sample for the targeted VNTR loci.
To calculate this statistic, the FBI applies the “product rule.” Use of the product rule in this context requires two assumptions about the statistical independence of allele matches: (1) that there is no greater or lesser likelihood that a person carrying one allele at a VNTR locus will also carry another particular allele at the same locus; and (2) that carrying one pair of alleles at a locus neither increases nor decreases the chance of carrying another particular pair at a different locus on a separate chromosome. If these assumptions *373are proper, then the product rule indicates that multiplying the population frequencies of all alleles detected in a DNA sample will yield an estimate of how common that DNA profile is in a given population. In this particular case, the FBI calculated that the chance that an unrelated individual chosen at random from the Caucasian population would have a DNA profile matching the forensic sample pattern is 1 in 50,000.
H. The Legal Standard of Admissibility
The sole issue we address in this appeal is the defendant’s contention that the trial court misapplied the legal standard for the admissibility of novel scientific evidence. Although both parties agree that the proper standard for determining the admissibility of scientific evidence is derived from Frye and Coolidge, they differ with regard to its proper formulation. We note that although we have recently adopted the New Hampshire Rules of Evidence, neither party has asked us to reconsider our use of the Frye standard.
Most courts that have considered the admissibility of novel scientific evidence have adopted the Frye test. See Jakobetz, 955 F.2d at 794 (describing Frye test as majority rule). The Frye court stated in its seminal decision:
“Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is difficult to define. Somewhere in this twilight zone the evidential force of the principle must be recognized, and while the courts will go a long way in admitting expert testimony deduced from a well-recognized scientific principle or discovery, the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs.”
Frye, 293 F. at 1014. “The Frye court assumed that general acceptance indicated reliability and that only reliable evidence should be admissible.” Jakobetz, 955 F.2d at 794. As one commentator noted:
“In effect, Frye envisions an evolutionary process leading to the admissibility of scientific evidence. A novel technique must pass through an ‘experimental’ stage in which it is scrutinized by the scientific community. Only after the technique has been tested successfully in this stage and has passed into the ‘demonstrable’ stage will it receive judicial recognition.”
*374P. Giannelli, The Admissibility of Novel Scientific Evidence: Frye v. United States, a Half-Century Later, 80 Colum. L. Rev. 1197, 1205 (1980).
Although the Frye standard has received substantial criticism, adherence to it: (1) permits disputes concerning scientific validity to be resolved by the relevant scientific community, United States v. Addison, 498 F.2d 741, 743-44 (D.C. Cir. 1974); People v. Barbara, 400 Mich. 352, 405, 255 N.W.2d 171, 194 (1977); (2) ensures that “a minimal reserve of experts exist who can critically examine the validity of a scientific determination in a particular case,” Addison, 498 F.2d at 744; (3) spares courts from the time-consuming and difficult task of repeatedly assessing the validity of innovative scientific techniques, Reed v. State, 283 Md. 374, 388, 391 A.2d 364, 371-72 (1978); and (4) “promote[s] a degree of uniformity of decision,” People v. Kelly, 17 Cal. 3d 24, 31, 549 P.2d 1240, 1244-45, 130 Cal. Rptr. 144, 148-49 (1976).
The defendant contends that the trial court applied a single-prong Frye test that was deficient for two reasons. First, he argues that the trial court failed to determine whether the particular technology employed by the FBI in performing its DNA analysis was generally accepted in the relevant scientific community. Second, he argues that the trial court erred by not assessing the reliability of the particular test results. The State contends that the trial court did consider the general acceptance of the techniques used in DNA profiling and that the reliability of the particular test results is a matter which affects the weight of the evidence, not its admissibility.
We adopted the Frye standard in Coolidge, 109 N.H. at 421-22,260 A.2d at 560-61. We stated the Frye test as follows: “[I]n order for the results of scientific tests to be admissible in evidence, the scientific principle involved ‘must be sufficiently established to have gained general acceptance in the particular field in which it belongs.’” Id. at 421, 260 A.2d at 560 (quoting Frye, 293 F. at 1014). In Coolidge, we examined the admissibility of evidence derived from neutron activation analysis and held that the Frye standard of general acceptance had been satisfied for some applications but not others. In applying Frye, we upheld the trial court’s exclusion of hair identification evidence developed by means of neutron activation analysis based on expert testimony that the tester’s methods would not be acceptable to scientists in the field. Id. at 420-22, 260 A.2d at 561. However, with regard to the comparison of particles vacuumed from the victim’s and defendant’s clothing, also derived from neu*375tron activation analysis, we stated that the trial court “could properly find that the test of particles produced an accurate analysis of the chemical elements which they contained, by means of procedures sufficiently accepted by scientists familiar with this limited field.” Id, at 422, 260 A.2d at 561 (emphasis added). Thus, although in Coolidge we did not plainly state the scope of our inquiry under Frye, it is clear that our analysis went beyond mere inquiry into the general acceptance of the theory underlying neutron activation analysis.
Generally, courts applying the Frye standard to determine the admissibility of DNA evidence have employed a two-prong test that requires both the theory and the techniques implementing the theory to be generally accepted in the relevant scientific community. See, e.g., United States v. Yee, 134 F.R.D. 161, 194 (N.D. Ohio 1991) (admissibility of DNA evidence is conditioned on general acceptance of principles and procedures); State v. Ford, 301 S.C. 485, 488, 392 S.E.2d 781, 783 (1990) (under Frye, admissibility of scientific evidence depends upon general acceptance of theory and technique); see also G. Lilly, An Introduction to the Law of Evidence 494 (2d ed. 1987) (in applying Frye, courts tend to apply standard of general acceptance to validity of scientific principle and process).
The defendant argues that the Frye test is a three-prong analysis similar to that used in People v. Castro, 545 N.Y.S.2d 985 (Sup. Ct. 1989). The Castro court acknowledged that New York follows the Frye test, id. at 986, but because of the complexity of DNA profiling evidence and the potential, powerful impact it might have on the jury, the court added a third prong as an additional hurdle to the admissibility of scientific 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 H. 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?”
Id. at 987. In applying its test, the Castro court found that the theory underlying DNA testing and the tests themselves met the Frye standard of admissibility. Id. at 999. The court, however, ultimately held that the inculpatory evidence derived from the tests was inad*376missible because the laboratory had failed to follow generally accepted techniques. Id.
The first two prongs of the Castro court’s test are firmly embedded in the Frye test. See Jakobetz, 955 F.2d at 794. The third prong, however, reaches beyond the requirements normally associated with Frye. We are aware of no court that has included the third prong as part of its Frye analysis. Even the Castro court acknowledged that its third prong involved a consideration of factors beyond the scope of the Frye test. See Castro, 545 N.Y.S.2d at 988 (first two prongs deal exclusively with Frye test and third prong involves issue of reliability of particular evidence); see also People v. Mohit, 579 N.Y.S.2d 990, 992 (County Ct. 1992) (third prong developed in Castro not properly part of Frye test).
Based on our consideration of how the Frye test is applied in other jurisdictions and our decision in Coolidge, we conclude that the admissibility of scientific evidence requires: (1) general acceptance in the relevant scientific community of the scientific theory or principle; and (2) general acceptance in the relevant scientific community of the techniques, experiments, or procedures applying that theory or principle. In our opinion, the third prong applied by the Castro court, as to whether the testing laboratory adhered to generally accepted techniques, addresses matters that properly go to either the admissibility or the weight to be given the evidence in a particular case, not admissibility under Frye. See Mohit, 579 N.Y.S.2d at 992 (concluding that third prong in Castro should go to weight of evidence, not its admissibility).
We now turn to an examination of the general acceptance of the theory and technology of DNA profiling techniques, including both RFLP analysis and population frequency calculation. The State urges us to uphold the trial court’s finding of general acceptance unless we decide that it was an abuse of discretion. However, whether a scientific theory and the technique used to implement it are generally accepted does not vary according to the circumstances of each case, and thus the determination of general acceptance is not a matter to be left to each trial judge’s individual discretion. See Reed, 283 Md. at 381, 391 A.2d at 367. Therefore, on appeal, we independently review the record and make our own determination of general acceptance without regard to the findings of the trial court. See Commonwealth v. Curnin, 409 Mass. 218, 223, 565 N.E.2d 440, 443 (1991); see also Giannelli, supra at 1222-23.
The majority of the jurisdictions that have ruled on this issue have found the DNA profiling theory and procedures for declaring a *377match to be generally accepted as reliable. See State v. Montalbo, 73 Haw. 130, 144-46, 828 P.2d 1274, 1283 (1992) (evidence derived from DNA testing admissible under Frye standard and Rules 702 and 703); State v. Brown, 470 N.W.2d 30, 32 (Iowa 1991) (finding that procedure was sufficiently reliable and met general test for admissibility); Smith v. Deppish, 248 Kan. 217, 239, 807 P.2d 144, 159 (1991) (trial court did not err in admitting DNA evidence because testing and RFLP analysis is recognized as reliable); State v. Davis, 814 S.W.2d 593, 602-03 (Mo. 1991) (finding reliability of procedures sufficiently established, and no abuse of discretion in admitting DNA evidence); State v. Pennington, 327 N.C. 89, 100, 393 S.E.2d 847, 854 (1990) (DNA testing method is reliable and trial court did not err in admitting DNA evidence); Ford, 301 S.C. at 488-90, 392 S.E.2d at 783-84 (DNA print testing and RFLP analysis recognized as reliable and gained general acceptance); State v. Wimberly, 467 N.W.2d 499, 505-06 (S.D. 1991) (holding DNA analysis meets Frye test of general acceptance); Spencer v. Commonwealth, 238 Va. 275, 290, 384 S.E.2d 775, 783 (1989) (undisputed evidence established DNA testing as generally accepted in scientific community), cert. denied, 493 U.S. 1036 (1990); State v. Woodall, 385 S.E.2d 253, 260 (W. Va. 1989) (finding DNA typing analysis generally accepted). While recognizing DNA testing as generally accepted, some courts have imposed limitations on the admissibility of population frequency statistics. See Caldwell v. State, 260 Ga. 278, 393 S.E.2d 436 (1990) (DNA evidence admissible, but only with conservative frequency estimate); Curnin, 409 Mass. at 224-27, 565 N.E.2d at 444-45 (finding method used by tester to calculate population frequency not generally accepted); State v. Schwartz, 447 N.W.2d 422, 428-29 (Minn. 1989) (evidence of match derived from RFLP analysis admissible under Frye standard, but accompanying statistics inadmissible); Mohit, 579 N.Y.S.2d at 995, 999 (finding FBI's method for conducting RFLP analysis and declaring match generally accepted, but population frequency estimate admissible only if most conservative method used). Generally, courts that have excluded DNA profiling evidence have done so because the particular testing laboratory failed to adhere to generally accepted techniques for obtaining relevant, reliable results, not because the theory or procedures were not generally accepted. See, e.g., Castro, 545 N.Y.S.2d at 999 (test results deemed inadmissible because laboratory failed to follow generally accepted techniques); Woodall, 385 S.E.2d at 260 (particular test results inadmissible under Rule 401 as inconclusive).
*378 A. General Acceptance of DNA Theory
The defendant does not contest the general acceptance of the theory underlying DNA profiling. As commentators have stated:
“There is nothing controversial about the theory underlying DNA typing. Indeed, this theory is so well-accepted that its accuracy is unlikely even to be raised as an issue in hearings on the admissibility of the new tests. . . . The theory has been repeatedly put to the test and has successfully predicted subsequent observations.”
Thompson & Ford, supra at 60-61. Based on our review of the record and the available literature, we also conclude that the theory underlying DNA profiling is generally accepted in the relevant scientific community. In future cases, a trial court may properly take judicial notice of its general acceptance and thus avoid relitigation on this issue.
B. General Acceptance of DNA Profiling Forensic Techniques
DNA profiling primarily involves the scientific disciplines of molecular biology and population genetics. The RFLP laboratory procedures that are used to determine whether there is a match between the sample taken from the suspect and the sample taken from the crime scene are largely drawn from the fields of molecular biology, biochemistry, and related fields. The significance of a declared match, as expressed by the probability that there is a coincidental match, is a matter of population and human population genetics. Accordingly, it is helpful to analyze the general acceptance of each procedure separately.
1. RFLP Analysis
At the pretrial Frye hearing, Dr. Adams testified to the general acceptance within the scientific community of forensic scientists of each step used in RFLP analysis individually and together. Drs. Daiger and Goldman each testified to the general acceptance in the field of molecular biology of the methods used in the FBI’s RFLP analysis. Dr. Shields, a defense expert, agreed with the State’s experts that the RFLP process employed by the FBI, except their use of a matching window, was generally accepted in the scientific community of molecular biologists. Dr. Nadeau, another defense expert, also agreed that the methods used in RFLP analysis are well accepted. In its order, the trial court noted that the RFLP process is used in thousands of laboratories worldwide for hundreds of different purposes and concluded that “[t]here does not appear to be any serious ques*379tion in the field of molecular biology that the RFLP process used in DNA profiling to measure or size the number of repeating fragments of the DNA is generally accepted.” (Emphasis added.)
The defendant’s primary challenge centers on the transfer of this science to the field of forensics, not the general acceptance of the RFLP techniques themselves. At the hearing, the State produced evidence that although the application of these methods in the field of forensics may be a “societal breakthrough,” it is not a “scientific breakthrough.” After reviewing the record, we agree with the analysis of the South Carolina Supreme Court:
“We recognize that the use of DNA analysis in forensic settings is a recent development. This type of analysis has been utilized for a number of years in diagnostic settings. Because the focus is different than in diagnostic settings, problems may exist that are unique to forensic DNA tests. For example, in forensic DNA testing, there is a higher probability that the sample may be contaminated by bacteria. Such problems, however, concern the reliability of the particular tests performed in a particular case. . . .”
Ford, 301 S.C. at 489, 392 S.E.2d at 783. As such, we agree with the trial court’s finding that the RFLP analysis, as outlined above, used to determine a match is a generally accepted technique in the scientific community.
The defendant raises three additional challenges to the general acceptance of RFLP analysis: (1) the FBI has chosen an improper matching window for determining whether there is a match between the defendant and the unknown sample; (2) the lack of objective criteria used in the FBI’s matching process in which the examiner conducts a two-step analysis of the autorad; and (3) the FBI’s environmental insult validation studies were insufficient. These issues all involve consideration of the reliability of particular test results and not whether the FBI’s techniques are generally accepted as capable of producing reliable results. As we noted above, such considerations normally go to either the admissibility or the weight to be given the evidence in a particular case, not admissibility under Frye. See Ford, 301 S.C. at 490, 392 S.E.2d at 784 (particular techniques used in specific test or reliability of test results may be impeached by expert testimony).
2. General Acceptance of Population Frequency Calculation
At the Frye hearing, Drs. Daiger and Goldman testified that, in the field of human population genetics, the methodology by which the *380FBI calculates population frequency estimates was generally accepted. Additionally, Dr. Conneally testified as to the general acceptance of the FBI’s fixed bin methodology for calculating population frequencies in the field of human population genetics and the FBI’s use of a randomly selected data base. In its pretrial order, the court concluded that “[t]he science of population frequency projections— be they the allele patterns of mice, fruit flies or humans—has been accepted for decades____[W]hat is important is that the FBI, as well as the commercial laboratories, do establish a population data base and follow widely accepted methods for making a population frequency calculation.” (Emphasis added.) Additionally, the court noted that the “statistical equation or formula has been in existence and used by scientists in the field for most of this century.”
In challenging the general acceptance of population frequency calculations, the defendant specifically attacks the reliability of the FBI’s data base. The defendant’s most important challenge is to the possible existence of population substructure in the Caucasian data base that the FBI used. Drs. Shields and Nadeau testified for the defense that they recognized the possibility of population substructure. Dr. Shields testified that due to substructure, the FBI cannot reliably use the product rule in their calculations, and that he was aware of population geneticists who disagreed with the FBI’s method of calculating population frequencies. Dr. Nadeau testified that population substructure would compromise the FBI’s method for calculating the probability of a random match.
Recently, the National Research Council (NRC), an organization administered jointly by the National Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine, appointed a committee to address the issues surrounding forensic DNA testing. In April 1992, the committee released a report entitled “DNA Technology in Forensic Science.” Although the NRC’s exhaustive report supports the general acceptance of most of the FBI’s methodology, its discussion of the debate regarding the existence of population substructure is particularly relevant to population frequency calculation.
As described above, the FBI applies the product rule by multiplying the individual allele frequencies together to calculate the frequency of the complete DNA pattern. The product rule is a basic statistical tool used in estimating the probability of a random match. According to the NRC report and the defendant’s expert witnesses, because the product rule is based on the assumption that each individual’s alleles constitute statistically independent evidence, its *381validity rests on the absence of population substructure. NATIONAL Research Council, DNA Technology in Forensic Science at 3-4, -6 (1992) [hereinafter NRC REPORT], Thus, the most important question underlying the validity of using the product rule is whether significant population substructure exists. Id. at 3-6.
On this issue, the report recognizes that a considerable debate exists among population geneticists. On the one hand, some population geneticists contend that population genetics studies show some sub-structuring within racial groups and that the absence of substructuring for any particular genetic marker cannot be predicted, but must be determined empirically. There are population geneticists, however, who acknowledge the possibility of population substructure, but argue that current data suggests that its effect on population frequency calculations is minimal. Although in its report the NRC does not commit itself to either side of this debate, it “assumes for the sake of discussion that population substructure may exist. . . .” Id.
In light of the conflicting expert testimony at the Frye hearing and the NRC’s recognition of considerable debate among population geneticists concerning the possibility of significant population substructure, we conclude that the FBI’s method for estimating population frequencies, which relies on the product rule, has not found general acceptance in the field of population genetics. See Commonwealth v. Lanigan, 413 Mass. 154, 162-63, 596 N.E.2d 311, 316 (1992) (finding FBI’s method for calculating frequency of defendant’s DNA profiles not generally accepted because of lively and current debate regarding existence of population substructure). Thus, we hold that the trial court’s decision to admit the population frequency estimates was error.
III. Conclusion
After considering the findings of the trial court and carefully reviewing the record, we hold that: (1) the theory underlying DNA profiling analysis is generally accepted in the relevant scientific community; (2) the technology that the FBI presently uses to conduct RFLP analysis and declare a match is generally accepted in the relevant scientific community as capable of producing reliable results; but (3) the statistical techniques that the FBI used to estimate population frequencies is not generally accepted among population and human population geneticists because of the debate concerning population substructure. A match is virtually meaningless without a sta*382tistical probability expressing the frequency with which a match could occur. NRC Report, supra at 3-1; see People v. Barney, 8 Cal. App. 4th 798, —, 10 Cal. Rptr. 2d 731, 742 (1992) (describing statistical calculation as “pivotal element” of DNA analysis). Thus, evidence of a match will not be admissible if it is not accompanied by a population frequency estimate that has been produced from a generally accepted method.
We are mindful that forensic DNA testing is an evolving science and that future discoveries or technological advances might lead us to reach a different conclusion than we have reáched today. Currently, two approaches offer great promise for addressing the issue of population substructure. The NRC recommends immediate empirical studies of ethnic subgroups to determine the extent of population substructure. Possibly, such studies will confirm that substructure does not exist or is minimal, thereby leading to general acceptance among population geneticists of the FBI’s method of calculating population frequencies. More important in the short term, perhaps, is the NRC’s suggested method for conservatively estimating population frequencies in order to account for population substructure. The report recommends using the “ceiling principle.”
The ceiling principle requires insertion of a “ceiling frequency,” or upper bound, for each allele at each locus when employing the product rule. NRC Report, supra at 3-10 to 3-11. The NRC urges that population geneticists conduct population studies of ethnic subgroups in order to provide for valid estimation of ceiling frequencies. The report describes this approach in detail.
“The ceiling principle yields the same frequency for a genotype, regardless of the suspect’s ethnic background, because the reported [ceiling] frequency represents a maximum for any possible ethnic heritage. Accordingly, the ethnic background of an individual suspect should be ignored in estimating the likelihood of a random match. The calculation is fair to suspects, because the estimated probabilities are likely to be conservative in their incriminating power.”
NRC Report, supra at 3-13.
The NRC asserts that the ceiling principle can account for any error caused by possible population substructure. Therefore, the admissibility of population frequency estimates does not necessarily await resolution of the population substructure issue, as long as the relevant scientific community generally accepts a method for calcu*383lating statistical probabilities. For instance, the State may be able to demonstrate general acceptance of the NRC’s recommended ceiling principle, which embraces the possibility of population substructure and thus yields a conservative estimate resolving all uncertainties in favor of the defendant. See Lanigan, 413 Mass, at 163, 596 N.E.2d at 316 (citing Cumin, 409 Mass, at 226-27, 565 N.E.2d at 445).
In light of our holding in this bifurcated appeal, we remand this case to the trial court. The trial court must conduct a hearing in order to determine whether the NRC’s recommended ceiling principle is a generally accepted technique. If the ceiling principle has gained general acceptance in the relevant scientific community, then the trial court must decide whether admission of the population frequency statistic in this case was harmless error. Resolution of the remaining questions raised on appeal are therefore stayed pending the trial court’s expedited determination of the remanded issues.
Reversed and remanded for further proceedings consistent with this opinion.
All concurred.