*842 Opinion
—The issue presented is one of first impression in this state: does DNA typing evidence meet the legal requirements for admissibility of novel scientific evidence and, if so, whether the basis for the calculation of statistical probability employed by the testing laboratory, Cellmark in this case, satisfied the foundation requirements of
People
v.
Collins
(1968)
Lynda Patricia Axell appeals from a judgment of conviction of first degree murder (Pen. Code, § 187, subd. (a)) and attempted robbery (Pen. Code, §§ 664/211) after court trial. She challenges the court’s ruling on admissibility of DNA testing results, alleges she was deprived of due process by admission of this evidence, and attacks her conviction as lacking substantial evidentiary support.
Following a preliminary hearing at which appellant was held to answer, the Ventura County Superior Court held lengthy hearings between March and August of 1989, to determine the admissibility of DNA testing results. The court ruled that the results of the DNA testing procedures, as well as statistics of probability concerning those results, were admissible.
The parties stipulated to a court trial on the basis of the preliminary hearing transcript, a tape recording of an interview between appellant’s father and Ventura Police Department officers, the transcript of all the proceedings in the superior court, and all exhibits introduced and stipulations entered into during those proceedings. The court found appellant not guilty of burglary, but guilty of first degree murder and attempted robbery.
Facts
The Offense
February 24, 1988, at approximately 8 a.m., Christine Kiloh, a cook at Top Hat Burger in Ventura, saw a person with dark hair three or four inches below the shoulder walk away from the Top Hat with head down, wearing a wide-brimmed hat and carrying an object under the arm. Upon entering the restaurant, she saw the body of George White, who customarily opened the restaurant in the morning, lying among boxes in a pool of blood. Blood was smeared and spattered in virtually every area of the building. Long hair was found throughout the premises including in the blood on the wall near the victim’s head. More unknown hairs were found on the victim’s trousers and on his body during autopsy.
*843 Several days later, the owner of the Top Hat, Charlotte Bell, discovered a few items missing from the restaurant. Although White usually carried a wallet, no wallet was found in his pockets at the murder scene, at his home, or in his truck.
Appellant, who had waist-length dark hair, worked in a business, The Party Place, near the Top Hat and was seen driving away from that area early on the morning of the murder. When interviewed by the police, she said she had gone to work at The Party Place about 4 a.m., worked until about 8 a.m., drove to her mother’s house, showered, went home, changed clothes, took a nap, awoke late, and finally returned to work around 9:30 a.m. She gave contradictory statements about whether her roommate Rhonda accompanied her back to work and whether she had ever been in the Top Hat. She denied involvement in the crime.
Douglas McCormick, nine years old, caught the bus to school in front of the Top Hat each day at approximately 7:40 a.m. and would talk to George White. On the morning of February 24, 1988, Douglas looked through the window of the Top Hat to see if George was inside and saw a person with grey hair and brownish skin, wearing a blue and white hat with “L.A.” on it, bending down, away from him. The person shooed Douglas away with a white rag. Several minutes later, Douglas looked in again but the person was gone.
Although he never saw the full face of the individual, Douglas thought the person he saw was a man in his 40’s. During cross-examination, he was shown a photograph of appellant but did not identify her as the person inside. He had previously picked another person’s photograph from a photographic lineup.
Edwin Jones, a Ventura County crime laboratory criminalist, compared hair samples of different persons, including appellant, with various hair samples from the crime scene. He eliminated the victim and other possible suspects as the source of the hairs found at the scene. Most of the hairs found ranged between 13 and 17 inches in length as did most of the hair samples known to have come from appellant. Appellant’s hair showed extreme variation in color apparent even without a microscope. The unknown hairs shared with appellant’s a dearth of follicular material adhering to the roots.
Many of the hairs found varied similarly in both diameter and cross-section, from nearly circular to oval, like that of appellant. Other similarities were shapes, pigment size, pigment distribution, lack of curl, and changes *844 along the lengths. Jones opined that the majority of the population would not have the same range of hair characteristics as appellant.
July 28, 1988, Cellmark Diagnostics, a testing laboratory in Germantown, Maryland, received from the district attorney’s investigator, whole bloodstains on cotton from the victim and appellant, and roots from 15 hairs recovered from the crime scene. The DNA was extracted from these materials, and Cellmark reported that the banding patterns obtained from the appellant’s whole bloodstain matched the DNA banding pattern obtained from the 15 hair roots found at the scene of the murder. Subsequently, Cellmark reported that the frequency of that DNA banding pattern in the Hispanic population is approximately 1 in 6 billion. Appellant is part Hispanic. Simply put, Cellmark’s analysis meant that the chance that anyone else but appellant left the unknown hairs at the scene of the crime is 6 billion to 1.
Appellant’s sister, Julie Burkette, in response to an untrue statement by Officer McKendry that appellant had confessed to the murder, said that she had overheard appellant tell someone at a family party on February 25,1988, that she had hit someone and taken their money. At the preliminary hearing she testified that appellant had never said that she had hit George, burglarized the Top Hat, or killed George White.
When officers told appellant’s father, Vernon Porcho, that they had taken her into custody for the killing and that she had confessed, Mr. Porcho told them that after the killing appellant told him that she went to the Top Hat to obtain money, that White entered and caught her, and she killed him. He said that his family disbelieved her admissions because of her chronic cocaine use. Her family had refused her request for money several days before the killing. At one point in the interview, he began to cry. Later, at the preliminary hearing, he changed his statement and testified that he had lied to the officers because he was mad at his daughter and at the police. He said appellant never told him that she killed George White.
John Adams, Julie Burkette’s then housemate and longtime friend of appellant, told police that appellant discussed the murder with appellant’s parents and that he overheard her say that she stabbed George White. He also said that Julie Burkette told him that appellant told her that she had hit and stabbed White.
Evidence Code Section 402 Hearing
A. The Nature of DNA
The scientific theory underlying DNA typing was explained at the
Kelly/ Frye
hearing. DNA, deoxyribonucleic acid, is a fundamental material which
*845
determines the genetic properties of all living things. All nucleated cells of every human being contain DNA. In 1952, scientists James Watson and Francis Crick won the Nobel Prize for their discovery of the structure of DNA. The molecule is composed of two parallel chain-like structures, a double helix, which has been described as a spiral staircase. The handrail and balustrade of the staircase is made up of repeated sequences of phosphate and deoxyribose sugar. Attached to the sugar links are four types of chemical bases—Adenine (A), Cytosine (C), Guanine (G), and Thymine (T). Pairs of these bases or “base pairs” form the steps of the spiral staircase. (See
People
v.
Castro
(1989)
In a single molecule of DNA, there are approximately three billion of these base pairs. Within an individual, the DNA from the nucleus of every cell is identical. Thus, the sequence of base pairs in a cell from a strand of hair will be identical to the sequence of base pairs found in a skin cell from the same individual. 1
Most of the DNA does not vary from person to person, thus creating such shared features as arms and legs. Other regions of DNA, however, vary distinctly from one person to another and it is these regions of the DNA molecule where the base pairs are arranged differently which make it possible to establish identity and differences between individuals. These variable regions of the DNA molecule are called “polymorphic loci” or “polymorphic sequences.” Every locus that is known has an official name and symbol assigned by the international Human Gene Mapping Workshop. At the time of the hearing, there were 1,500 polymorphic DNA loci known.
The human DNA molecule has 23 pairs of chromosomes, 22 from each parent and 2 sex-typing chromosomes, 1 chromosome in each pair inherited from each parent. Each chromosome has thousands of genes and each gene is located on a particular site on the chromosome. 2 Each gene or segment of DNA material that produces a trait is called an allele. Thus, an allele is one *846 of several forms of a gene, occupying a given locus on the chromosome. In variable regions of DNA, a pair of chromosomes is distinguished by different alleles—one from the mother and one from the father. If the mother and father each contribute the same general information for a particular trait, that locus is homozygous; when the genetic information differs, the locus is heterozygous.
B. DNA “Fingerprinting”
“Restriction enzymes” are able to recognize a specific sequence of six base pairs along a strand of DNA. When the enzyme recognizes this particular sequence it cuts the DNA at that location. These enzymes enable the researcher to cut the long DNA molecule into shorter fragments which can be used for many research purposes and for DNA “fingerprinting” or typing. The two strands of the DNA chain or staircase can also be separated or “unzipped.”
DNA typing, or restriction fragment-length polymorphism (RFLP) analysis, involves a number of steps: (1) extraction and purification of the DNA; (2) fragmentation by restriction enzymes; (3) gel electrophoresis in which a positive electrical charge to the bottom of an agarose gel on which a DNA sample is placed causes the DNA to move through the gel from the negative to the positive charge; (4) Southern blotting in which the gel and DNA in it are transferred to a nylon membrane for easier handling; (5) hybridization in which the DNA pattern unique to the individual is identified by use of radioactively tagged probes, “unzipped” DNA segments of a known length and sequence, designed to seek out a predetermined locus in a polymorphic region of the DNA and band with a like segment of DNA; and (6) autoradiography in which a film is developed on top of the nylon membrane, revealing the location of the DNA by bands on the X-ray film, called an autoradiogram or autorad. Use of a single probe produces two bands on the autorad. Thus, running four different probes at the same time results in eight bands.
The autorads must be interpreted and the bands produced by the migration of DNA in the gel in different lanes examined to ascertain if they match. Essentially the bands on the autorad from the victim’s, suspect’s, and crime scene evidence samples are “eyeballed” to see if they match within a certain measurement. If a match is declared, the likelihood that a match is unique must be determined. A match is said to occur if the sizes and number of the detected DNA fragments in various lanes are indistinguishable within a permissible degree of error. To calculate the permissible degree of error, *847 Cellmark uses “resolution limits” as a unit of measurement to ascertain the “bin” or frequency at which an allele occurs in the population data base.
C, Statistical Analysis
To make a statistical evaluation of the data obtained from a DNA typing, it is necessary to know how frequently in the population a band of a certain size will be found, a question answered according to the principles of population genetics. Each probe recognizes a pair of bands—one from each parent. The probability of the combination of two particular bands recognized by one of the probes is calculated by multiplying the product of the frequencies of the two bands by two. The probability of the band patterns from all four loci is determined by multiplying the products from all four loci. This is known as the “product” or multiplication rule. 3
The validity of this procedure presupposes that each fact observed, and entering into the calculation, is random and independent of the others, or adjustments are made for deviations from conditions known as “Hardy-Weinberg equilibrium” and “linkage equilibrium.”
(People
v.
Wesley, supra,
As explained in
People
v.
Castro, supra,
D. The Experts
At the hearing, the People presented four expert witnesses in their casein-chief and two in rebuttal. Doctor Richard Roberts, a molecular biologist, is assistant director for research at Cold Spring Harbor Laboratory in New York, one of the foremost laboratories for the study of genetics and molecular biology.
Doctor Roberts, considered the world’s leading authority on enzyme restriction, testified that he examined the autorads of Cellmark labeled *849 “Linda Axell” and “hair roots” and, in his opinion, the bands for the two had matched. He opined that the autorads indicated that the procedures had been properly conducted. After reviewing the protocols used by Cellmark labs and laboratory notes in this case, Doctor Roberts was of the opinion that the procedures set forth in the protocol were recognized as reliable in the scientific community, and, as performed in this case, conformed to the protocols.
Doctor Kenneth Kidd, a professor of human genetics, psychiatry, and biology at the Yale University School of Medicine, is coauthor of The Genetics of Human Populations, one of the primary reference works in the field of human population genetics. Doctor Kidd has published over 200 papers in his field, and is active in the international Human Gene Mapping Workshop, which compiles all current information about the distribution of genes on human chromosomes.
Doctor Kidd testified that he had examined Cellmark’s protocol and lab notes on this case. In his opinion, the protocols contained standard procedures for performing the RFLP process and the technician who tested the DNA in this case correctly followed the procedures outlined in the protocols. He also believed that Cellmark uses an acceptable data base and acceptable procedures for calculating the frequency of bands. As he encounters errors in the RFLP process in his own laboratory, he opined that it is essentially impossible for these errors to cause patterns which are different to look alike. He found no evidence of any technical errors in the autorads from this case.
Doctor Robin Cotton, manager of research and development for Cellmark Diagnostics, has a doctorate in molecular biology and biochemistry and has been a postdoctoral fellow in biochemistry. Doctor Cotton evaluates the protocols and designs the study for the data base at Cellmark which specializes in using DNA typing for identity purposes. She is familiar with the standard operating procedures used at Cellmark and reviewed the notes of Lois Tonelli, the technician who actually performed the procedures. She concluded that Ms. Tonelli had followed the procedures set up within the laboratory and that the methods used by Cellmark to calculate frequencies of banding patterns is one which is accepted in the scientific community.
Doctor George Sensabaugh, Jr., a forensic biologist and biochemical geneticist and professor of forensic science and forensic biology at the School of Public Health at the University of California, Berkeley, is an associate editor of the Journal of the Forensic Science Society and an editorial board member of the Journal of Forensic Sciences.
*850 Doctor Sensabaugh testified that, in his opinion, the idea that DNA could be used in the analysis of forensic evidence is generally accepted in the knowledgeable forensic community.
The defense produced testimony of four experts who opined that DNA typing evidence is not reliable and is not generally accepted for the purposes of forensic identification and three experts who expressed opinions that the data base used by Cellmark Diagnostics is inappropriate and unacceptable for calculating statistical probabilities of a DNA “match.”
Doctor Diane Lavett, a professor of genetics at State University of New York and at Emory University in Atlanta, Georgia, stated that she had reviewed the protocols of Cellmark as well as the laboratory notes from this case and found that Cellmark employs inadequate controls in their procedures, and that Cellmark had deviated from their protocols in performing the test in this case. She also opined that statistics reported by Cell mark are unreliable in that the data base appears to deviate from Hardy-Weinberg equilibrium. Doctor Lavett did not believe that DNA typing had been adequately considered by the scientific community or generally accepted.
Doctor Randall Libby, an associate professor of molecular genetics at University of Washington, opined that although RFLP technology is accepted as a reliable technique in research laboratories, the use of RFLP technology for criminal identification of forensic samples is not necessarily accepted as reliable in the scientific community. Particularly, he thought there was not a sufficient body of research or literature to determine the likelihood or unlikelihood of false positives under forensic conditions. He reviewed Cellmark’s protocols and found them lacking in appropriate controls and that the technician failed to follow the correct laboratory procedures set out in the protocols.
Doctor Simon Ford, an associate research professor of environmental molecular genetics at the University of California, Irvine, testified that he reviewed Cellmark’s protocols as well as over 100 autoradiograms. Although he agreed that RFLP analysis was an accepted technology in research laboratories, he believed that forensic use of the technology suffered from problems not encountered in research which render those techniques unreliable and unacceptable for purposes of identification from crime scene samples.
Doctor William C. Thompson, a professor in the department of social ecology at University of California, Irvine, conducted two surveys in 1988 of crime laboratory directors and molecular biologists and molecular geneticists regarding DNA testing and its application to forensic cases. Doctor *851 Thompson opined that, based upon the surveys, there is not a consensus in the scientific community that DNA typing is ready for forensic use.
Doctor Lawrence Mueller, an associate professor in the department of ecology and evolutionary biology at University of California, Irvine, analyzed tire data from Cellmark’s Hispanic data base and concluded that the data base did not conform to Hardy-Weinberg and linkage equilibrium.
Doctor Charles Taylor, a professor of biology at University of California, Los Angeles, also opined that Cellmark’s Hispanic data base was not appropriate for calculating uniqueness of appellant’s banding patterns because the data base did not conform to Hardy-Weinberg equilibrium. He also believed that the population sample was not representative of appellant’s population subgroup, making her patterns appear more unique than they were.
Doctor Seymour Geisser, professor of statistics and director of the school of statistics at the University of Minnesota did not believe the Hispanic population sample, drawn from a single blood bank in Los Angeles, was a random sample of the Hispanic population. He thought that the size of the sample used by Cellmark to determine relative frequence of band patterns is inadequate, and should consist of 5,000 to 10,000 persons. He also stated that no data had been presented to validate either the bands within a probe are independent of each other or that the probes are independent of each other. Without this independence, he stated, any statistical computations using the product rule are meaningless.
The prosecution presented Doctor Patrick Conneally and Doctor Lisa Forman as rebuttal witnesses: Doctor Conneally, distinguished professor of medical genetics and neurology with the University of Indiana Medical Center, is board certified in human genetics, director of the American Society of Human Genetics and secretary of the American Board of Medical Genetics. He received his doctoral degrees from the University of Wisconsin in statistics and human genetics, and, at the time he testified, was about to receive a Doctor of Science Honoris Causa from Trinity College in Dublin, Ireland. He is a member of the Human Genome Organization, which maps human DNA, serves on editorial boards of several publications in the field of genetics, and lists almost 300 publications in his curriculum vitae.
Doctor Conneally explained that the occurrence of single-handed RFLP patterns, which led Doctors Mueller and Taylor to conclude that Cellmark’s data base did not conform to Hardy-Weinberg equilibrium, could have three possible causes: (1) there are small bands which are not being observed; (2) there are really two bands which are so close together that they cannot be *852 distinguished from one another; or (3) the population is not in Hardy-Weinberg equilibrium. In his opinion, deviation from Hardy-Weinberg equilibrium was the least logical of the three explanations because this would indicate that the population sample was “substructured,” i.e., mated within certain subgroups such as race or ethnic group, and in his opinion United States populations do not exhibit substructuring.
He also opined that the probes used by Cellmark are in linkage equilibrium because they are on different chromosomes, or far enough apart to be transmitted independently. He stated that the data base for Hispanics used by Cellmark in this case is large enough to obtain reasonable probabilities.
Doctor Lisa Forman, a population geneticist with Cellmark with a doctorate in biological anthropology from New York University, is a guest researcher with the National Cancer Institute at the National Institutes of Health. She stated that Cellmark is presently in the midst of analyzing their data base to determine why it appears to deviate from Hardy-Weinberg equilibrium, and that it appears that there are small bands which are not being visualized in the data base autorads, thus creating the appearance of an excess of single-band patterns or homozygosity. Doctor Forman stated that preliminary analysis has shown that a significant number of the single-band patterns actually have two bands. She stated that she does not know with certainty that the data bases used by Cellmark are in Hardy-Weinberg equilibrium, although she expects them to be and that even where a data base is not in Hardy-Weinberg equilibrium, one can still use the data to generate meaningful probabilities.
Doctor Forman stated that obtaining samples of blood from blood banks is a standard way of data basing which has been used for blood enzyme systems over the years and that the samples provided by the Los Angeles Red Cross Blood Bank provided an appropriate data base.
The trial court, in ruling that the DNA test results were admissible, made the following findings: (1) the DNA identification procedure is an application of restriction fragment length polymorphisms known as RFLP, or otherwise known as RFLP technology; (2) RFLP technology is widely used and is a standard and accepted technique in the relevant scientific communities of molecular biologists and human geneticists; (3) the probes used by Cellmark Diagnostics recognize hypervariable regions on the human genome and are thus appropriate for use in identifying individuals according to a consensus of the relevant scientific communities of molecular biologists, human geneticists and forensic scientists; (4) the use of RFLP technology to aid in the identification of sources of forensic samples is widely accepted by the relevant scientific community of forensic scientists; (5) the use of RFLP *853 technology to eliminate a person as a possible source of forensic samples is widely accepted in the relevant scientific community of forensic scientists; (6) the calculation of frequencies of genetic types or genotypes in human populations for the use of data bases made up of samples from those populations has long been accepted practice in the relevant scientific communities of human geneticists and forensic scientists and has been accepted by the courts for many years.
The court found that issues regarding the size and accuracy of any given data base and the accuracy of calculated frequencies of genetic types affect the weight to be given to the probability testified to by the proponents of those figures, and that although the opinion concerning the population frequency is not exacting, the probative value is “very substantial nonetheless” whether a figure such as one in six billion is off by three or four billion. The court found the most critical factor in determining whether the prosecution’s burden has been met is no likelihood of a false match or a false positive.
Discussion
1. Admissibility of DNA Typing Evidence
a. Standard of Review
Appellant contends that the DNA typing evidence failed to meet the
Kelly/Frye
rule. In
People
v.
Kelly
(1976)
Under the
Kelly/Frye
rule, the proponent must establish the reliability of the method, usually by expert testimony, that the witness furnishing the testimony is properly qualified as an expert to give an opinion on the subject, and that correct scientific procedures were used in the particular case.
(Ibid.; People
v.
Kelly, supra,
Additionally, the witness must have sufficient academic and professional credentials to understand both the scientific principles involved and any difference of view concerning their reliability, and must be impartial.
(People
v.
Brown, supra,
*854
Kelly/Frye
does not demand that the court decide whether the procedure is reliable as a matter of scientific fact: the court merely determines from the professional literature and expert testimony whether or not the new scientific technique is accepted as reliable in the relevant scientific community and whether “ ‘scientists significant either in number or expertise publicly oppose [a technique] as unreliable.’ ”
(People
v.
Brown, supra,
The court may also consider decisions from other jurisdictions and relevant scientific literature in deciding whether a technique is generally accepted.
(People
v.
Morris
(1988)
b. Relevance of Decisions in Other Jurisdictions
Appellant concedes that DNA evidence has been admitted in trial courts in a number of jurisdictions, but argues that this fact should have little precedential effect because courts should not substitute the frequency of the use of a novel technique or its admission at trial for an analysis of its general acceptance. (See Hoeffel, The Dark Side of DNA Profiling: Unreliable Scientific Evidence Meets the Criminal Defendant (1990) 42 Stan.L.Rev. 465.)
Appellant points out that in several of the court decisions concerning admissibility of DNA (e.g.,
Andrews
v.
State
(Fla.App. 1988)
Appellant states that the only decision to date from a state Supreme Court,
State
v.
Schwartz
(Minn. 1989)
Appellant suggests that Schwartz is applicable here since she raises the same arguments and that law review articles (including one written by two of the defense experts, Thompson and Ford) conclude that DNA typing evidence has been introduced into the courtroom before it has received adequate scrutiny in the scientific community. (See Hoeffel, The Dark Side of DNA Profiling: Unreliable Scientific Evidence Meets the Criminal Defendant, op. cit. supra, and Jonakait, When Blood is Their Argument: Probabilities in Criminal Cases, Genetic Markers, and, Once Again, Bayes’ Theorem (1983) U.I11. L.Rev. 369.)
In
State
v.
Schwartz, supra,
*856
In
People
v.
Castro, supra,
Credible testimony here, as well as review of the above cited decisions, reveals that the steps of DNA fingerprinting involve scientific principles and technology all of which have gained general acceptance in the scientific field in which they belong. (See also
Cobey
v.
State, supra,
In
State
v.
Schwartz, supra,
Appellant apparently concedes that DNA typing may have attained general acceptance for research and diagnostic purposes but asserts that it is not reliable for forensic purposes. 8 She also attacks the credentials of the *857 prosecution experts to testify on this issue. Additionally, appellant contends that the application of RFLP analysis differs in these fundamental ways: (a) the degree of technical difficulty encountered in interpreting the results of forensic tests; (b) the nature of forensic samples; and (c) the lack of standards and regulations for forensic laboratories. She points out that forensic evidence samples, unlike those in research and medical diagnosis, are extremely limited in quantity and subject to contamination and degradation. The conclusion that a suspect matches a forensic sample is typically based on a single result which is not replicated.
Appellant also asserts that defense experts Doctor Ford and Doctor Libby discussed interpretational problems which arise as a result of contamination or degradation of the DNA tested and that research in that area is gestational at best. She notes that the DNA testing of disease genes involves a discrete allele system comprising of only a specific limited number of genes. Forensic testing for DNA, to the contrary, involves a continuous allele system, i.e., polymorphic groups containing unknown numbers of genes. She attacks the “total absence” of published research by Cellmark for review by defense experts and points to criticism on this issue in the scientific literature as well as case opinions. (See, e.g., Lander,
DNA Fingerprinting On Trial,
Nature (June 15, 1989);
State
v.
Schwartz, supra,
447 N.W.2d at pp. 426-427;
Com.
v.
Curnin
(1991)
We uphold the trial court’s finding that prosecution experts were duly qualified to testify. The search for scientific consensus must be from within “the particular field in which it belongs.”
(People
v.
Reilly, supra,
Here, the trial court heard the opinion of a wider spectrum of scientists than simply molecular biologists and population geneticists. Evidence disclosed that where forensic scientists, molecular biologists, population geneticists, attorneys, and government policymakers gathered at a multidisciplinary conference in December 1988, they concluded that DNA typing was usable in the analysis of forensic evidence.
Doctors Roberts, Kidd and Sensabaugh, all recognized as experts in their fields, testified that RFLP technology is generally accepted in the relevant scientific communities. Doctor Roberts is the world’s leading expert on *858 restriction enzymes and Doctor Kidd is a well-known population geneticist who the court found “probably the foremost expert” in the United States in this field. (Appellant does not challenge Doctor Sensabaugh’s credentials.) Doctors Roberts, Kidd, and Sensabaugh have testified in favor of admission of DNA typing in cases in other jurisdictions and were found qualified to testify.
Doctor Roberts explained that DNA typing is the application of a series of techniques used since 1974. The main difference between diagnostic and forensic use is that of the limited sample size and possible contamination in the forensic use. However, Doctor Roberts’s laboratory, which does not do forensic work, has extracted DNA from samples contaminated by dirt or dust. Moreover, serious contamination or degradation would likely result in inability to extract sufficient DNA to test rather than distorted results or false positives.
Doctor Kidd explained that medical diagnosis in using the technique requires a degree of reliability as high or higher than that required for forensic application since, in a forensic case, a “false negative” is acceptable as it benefits the defendant whereas an erroneous failure to identify a positive is not acceptable in medical diagnosis because the subject may die.
Doctor Kidd also testified that the probes used by Cellmark—MSI, MS31, g3 and MS43—have been published by the Human Gene Mapping Workshop and their inclusion in the workshop signifies the scientific community’s acceptance of the fact that the region of DNA identified by the probes are polymorphic. 9
Doctor Sensabaugh, a forensic biologist, testified that in his opinion, RFLP typing of DNA is generally accepted in the knowledgeable forensic community. Many articles have appeared in scientific literature which describe the forensic use of DNA and DNA has been the subject of papers presented at conferences since 1985. No one had ever presented a paper for peer review suggesting that RFLP or DNA typing is not scientifically valid or that problems exist which should preclude its use. 10
Additionally, the Congressional Office of Technology Assessment has found that DNA tests are valid and reliable in forensics when performed and
*859
analyzed properly by skilled personnel.
(U.S.
v.
Two Bulls
(8th Cir. 1990)
Appellant criticizes the partiality of Doctors Cotton and Foreman who are employed by Cellmark. We disagree with appellant that the record is unclear concerning the manner in which the court received their testimony. The trial court specifically noted that these witnesses worked for Cellmark and that it was considering their opinion only to understand more clearly Cellmark’s practices and methodology.
The trial court has wide latitude in determining the qualifications of an expert; once the court acts within its discretion and finds the witness qualified, the weight of the testimony is for the trier of fact.
(People
v.
Smith, supra,
Appellant asserts that Doctors Roberts, Kidd and Conneally are “so far removed” from relevant scientific community of forensic scientists that their opinions regarding the general acceptance of forensic DNA testing is of little value. We find this statement curious since the defense witnesses had less experience than prosecution witnesses in the actual procedures of testing human DNA in a forensic setting and none was a forensic scientist. Defense witnesses recognized the reputations of prosecution witnesses such as Roberts and Kidd as experts in their fields.
11
Moreover, scientists experienced in the field may testify “ ‘if presented with the information they need to fill the gaps in their own knowledge and experience.’ ”
(People
v.
Morris, supra,
Unlike in many cases in other jurisdictions excluding the evidence, the prosecution experts here explained perceived inadequacies pointed out by defense experts and successfully rebutted defense challenges. The trial court did not find Doctor Lavett to be a particularly credible witness as she propounded theories which she conceded were not generally accepted, was not trained as a molecular biologist and had never done RFLP analysis. Similarly, other defense witnesses raised hypothetical problems admittedly not likely to occur.
*860 According to the prosecution experts, it is extremely unlikely if the procedures described in Cellmark’s protocol are followed that an erroneous match could result from contamination or degradation of the DNA sample, from incomplete or partial digestion by the restriction enzyme, from star activity (where restriction enzyme recognizes additional sites), or aberrant band mobility, as these phenomena would be detectable from the results if not at an earlier stage or by appearance of the autorad. Even if it were theoretically possible to result in an erroneous match, the statistical likelihood of an artificial match at all eight bands is extraordinarily low.
Similarly, if the samples were too small, it would not be possible to detect a signal from it. If the DNA were degraded, the signal produced would be weakened to the point that no genetic pattern could be discerned. It is more likely that any threshold problem during RFLP analysis would make either an uninterpretable or inconclusive result, or the erroneous exclusion of a person as the source of the forensic sample, all “errors” beneficial to the suspect. 12 Here, Doctors Roberts and Kidd found the autorads and test results to be of high quality, first-rate science, and publishable. Moreover, there was no evidence of contamination, degradation, partial digestion, band shifting, or any other problem that might have affected the results.
Thus, the defense witnesses’ testimony on the issue of general acceptance did not undermine the validity of the trial court’s determination that forensic use of RFLP analysis is generally accepted in the relevant scientific community. “ ‘A determination of reliability cannot rest solely on a process of “counting (scientific) noses” ’ ”
(U.S.
v.
Jakobetz, supra,
c. Procedures Used by Cellmark
Similar challenges were raised to the admission of DNA typing by the FBI in
U.S.
v.
Jakobetz, supra,
*861
Jakobetz
appears to reflect the reoccurring refrain throughout cases on DNA typing in other jurisdictions: except for the Lifecodes’ debacle in
People
v.
Castro, supra,
As appellant points out, forensic laboratories performing DNA typing are not required to conform to any standards or guidelines to ensure reliability. The American Society of Crime Laboratory Directors (ASCLD) Guidelines for Forensic Laboratory Management and CACLD’s Report of the Symposium on Forensic Serology—1987 propose recommended practices for the validation and implementation of new methods in forensic laboratories, including seven guidelines or criteria to be met by a DNA testing laboratory in order for their method to be judged acceptable by the forensic scientific community. The FBI also promulgated a list of guidelines before DNA typing techniques will be certified for use as case evidence.
Doctor Cotton of Cellmark admitted that, although her company substantially complied with most of the guidelines, it had not disclosed the procedure used for typing, or made available probes to other laboratories for the purpose of replicating validation studies or independent re-analysis of the evidence. Further, Cellmark has failed to publish the results of any experimental studies conducted on either contaminated or aged stains. These lacunae were considered important in the Minnesota Supreme Court’s decision in
State
v.
Schwartz, supra,
Appellant also criticizes Cellmark’s lack of performing a mixing experiment in which a 50:50 mixture of two samples are mixed for marker DNA to test the source of each sample. This practice is a method of controlling human error which might result in a “false positive.” Cellmark had such an error in a blind proficiency test resulting in a 2 percent error ratio which, appellant asserts, is not acceptable.
Respondent asserts that any issue pertaining to whether the procedures were applied correctly in a given test is relevant not to admissibility but only to the weight and is, therefore, for the trier of fact to determine. (See
People
v.
Farmer
(1989)
We agree with appellant that the brief statement in
People
v.
Farmer, supra,
Due to the complexity of the DNA multisystem identification tests and the powerful impact that this evidence may have on a jury, satisfying
Frye
alone is insufficient to place this type of evidence before a jury without a preliminary critical examination of the actual testing procedures performed.
(People
v.
Castro, supra,
We hold that the procedures used in this case were the ones generally accepted as reliable in the scientific community. That defense witnesses found fault with particular aspects of the test or thought the protocol could have been improved did not invalidate the testimony of prosecution witnesses, especially since some of the improvements suggested, such as the ASCLD and CACLD guidelines, are not mandatory and suggestions such as the “mixing” experiment were not shown to be generally acceptable to forensic scientists.
*863
Similarly, that Cellmark’s procedure led to a false positive in a blind study performed for CACLD does not invalidate the procedure. (See
State
v.
Pennell, supra,
We hold that the trial court properly considered the views of a typical cross-section of the scientific community, including representatives of those who oppose or question the new technology, and that the prosecution’s experts were generally better qualified and more knowledgeable in the field.
d. Cellmark’s Criteria for Assigning Frequencies
Appellant challenges Cellmark’s procedure of determining the statistical probability of a random match by assigning a “frequency” to each individual band, and then applying the product rule to those individual frequencies. The individual frequencies are assigned by comparing the bands recognized by a particular probe to a “data point” on the data base. 13 A “bin” or frame is made around that data point, which includes all the bands within a single resolution on either side of the data point. Resolution limits are designed to account for variability in the measurement of bands and to provide quantitative means of determining a match. The number of bands within the bin is then said to represent the frequency of the questioned band within the population. 14
When comparing bands in samples run on the same gel, any bands within one resolution limit of each other are declared a match. When bands in samples run on different gels are compared, bands lying within two resolutions of each other are called a match to account for inconsistencies in the running of separate gels.
Appellant asserts that Cellmark made no allowance for the fact that the data base samples were likely run on many different gels. Consequently, she argues, many bands which may have been the same size as appellant’s were probably not included in the one resolution bin due to fluctuations from gel *864 to gel. She asserts that in assigning a bin size in the data base for the purpose of determining the frequency of the band, more than one resolution should be used in order to account for differing gel conditions in the data base gels which affect the measurement occurring on different autorads for the data base. 15 She contends that had Cellmark used a bin of two resolutions for purposes of assigning frequencies to her bands, the statistical probability of a match would have been significantly higher.
Doctor Forman stated that the bin one uses to match must be the same as the one used for frequencies, and if it is not, at least the one for frequencies must be greater than the one used to match. Doctor Forman explained that even though all of the data base samples are not run on a single gel, Cellmark uses a single resolution limit which was determined to be a reasonable bin size to compare samples with a single gel to see if they matched. It is considered to be a strict matching criteria, is a consistent method, and is fair, but not as conservative as having allele frequencies in two bin sizes.
Doctor Forman stated that if one uses a greater bin size, the overall probability, instead of 1 in 6 billion, might be 1 in 3 or 4 million. However, she noted that the total figure 1 in 6 billion or any other number, even using standard deviations instead of resolution limits, would still be approximate; it is the magnitude of the number that is important whether it is 1 in 6 billion or 1 in 1 million.
Doctor Kidd testified that Cellmark takes into account two margins of error in its procedure—measurement error and sampling error. Cellmark’s use of resolution limits was reasonably conservative and its method of using relatively large bins tends to cause a particular banding pattern to appear to be less rare than it actually is, which works to the defendant’s favor. Doctor Conneally also testified that Cellmark’s binning methods are acceptable and that its criteria for obtaining gene frequencies are identical to those for matching them.
Since expert testimony established that Cellmark takes into account a margin for error in its measurement, whether it could or did change its measuring procedures to make them more accurate would appear to go to the weight of tiie evidence more than its admissibility.
Appellant’s challenge to the subjectivity of interpreting a match does not invalidate the procedure. Doctor Sensabaugh testified that interpretation of
*865
bands on an autorad is fairly straightforward and involves a minimal amount of subjective analysis. More than nine out of ten times, two independent reviewers will reach the same conclusion. In this case, two Cellmark technicians independently agreed that appellant’s DNA sample matched. It is the laboratory analyst’s competence, experience, and training that are important.
(People
v.
Reilly, supra,
e. Foundation for Calculating Statistical Probability
Appellant asserts that Cellmark’s calculation of the statistical probability of a match is subject to the same infirmities discussed in
People
v.
Collins, supra,
Here defense witnesses testified that calculation of the probability of a random match in a population depends upon four major assumptions: (1) the correct population has been identified; (2) the population sample is large enough that the observed frequencies accurately represent the true population frequencies; (3) the sample is truly random; (4) the population is homogeneously mixed, in the technical sense that each locus is in Hardy-Weinberg equilibrium and the loci are together in linkage equilibrium.
Appellant asserts that Cellmark failed to identify the correct population with which to identify her. The population to which her alleles were compared was “Hispanic” which is described as one of Spanish surnames or of Spanish descent and could include blacks, whites, Filipinos, Mexicans, Spanish, and any number of subpopulations of South, Central or North America.
Doctor Taylor said the data base fails to include individuals of native American or South American ancestry and that single locus genotypes for appellant are underrepresented in the data base. Doctor Geisser thought the population sample for which the allele frequency have been derived is not large enough to accurately report the true population frequency. However, *866 appellant and her family characterized her background as “Hispanic,” which is a social or geographic term rather than genetic. Since an Hispanic data base may include those of American Indian ancestry, use of an “Hispanic” data base from Southern California was proper.
Doctor Forman testified that she spoke with the person in charge of sending the blood samples from Los Angeles and assured herself that he was familiar with statistics and understood the importance of random sampling. Doctor Kidd also opined that the data base was representative of Hispanics in Southern California. Doctor Forman acknowledged that there appeared to be an excess of single band patterns in the Hispanic data base. However, she testified that she agreed with many experts that calculations can be derived from the data presently available, regardless of their conformance to Hardy-Weinberg equilibrium. She also stated that, based upon discussions with Doctor Kidd, Doctor Conneally, and Doctor Eric Lander about the possible deviation of genotype predictions based on the likelihood of linkage disequilibrium, Cellmark can rely on information in the Hispanic data base to calculate gene frequencies without knowing whether or not there is linkage disequilibrium. She also disputed Doctors Mueller and Taylor because they did not completely analyze their data set.
“[B]oth California and the majority of other jurisdictions have traditionally admitted statistical blood-group evidence of this kind in criminal cases, even where it simply includes the accused within the class of possible donors.”
(People
v.
Brown, supra,
Some jurisdictions, as respondent points out, have considered that questions concerning contamination of a sample, chain of custody, reliability of particular results, as well as the size or ratio of the population frequency and statistical probabilities relate to the weight of the evidence and not its admissibility. (See, e.g.,
State
v.
Pennington, supra,
Two issues emerge: (1) are the methods used in calculating the statistical probabilities ones scientifically accepted as generally reliable in the particular field and (2) is there evidentiary support for the particular application? We find that since a match between two DNA samples means little without data on probability, the calculation of statistical probability is an
*867
integral part of the process and the underlying method of arriving at that calculation must pass muster under
Kelly/Frye.
However, the size or ratio of the population frequency is a matter of weight rather than admissibility. As in
People
v.
Castro, supra,
In
Com.
v.
Curnin, supra,
In
U.S.
v.
Jakobetz, supra,
In
Jakobetz,
as here, Doctor Kidd testified that from looking at data from many subgroups, i.e., Irish, Swedes, Amish, all have “very small differences” in allele frequencies. (
Doctor Kidd testified here that the data base was adequate and acceptable within the scientific community. Moreover, appellant’s loci here each displayed a double-band or heterozygous pattern. He opined that the sample is representative of Hispanics in Southern California, the relevant population
*868
group. Other courts have recognized that conservative or reduced calculations such as used by Cellmark may correct any Hardy-Weinberg deviation problems. (See
Caldwell
v.
State, supra,
Doctor Conneally testified that it is standard procedure to use blood banks and that there are only three inbred populations in the United States—the Mennonites, Amish and Hutterites. Doctor Mueller admitted that there was no evidence that Los Angeles Hispanics were inbred. Cellmark uses a separate data base for each of the four probes and the prosecution experts testified that population samples of 272 and 297 were adequate. (See also
Andrews
v.
State, supra,
Concerning linkage equilibrium, the relative independence of two genes in the population, Doctor Kidd testified that while two of the four single-locus probes used by Cellmark in this case showed slight linkage, the loci were approaching independence. Doctor Conneally explained that although two of the four loci identified by Cellmark’s probes did occupy the same chromosome, the alleles were so far apart as to be transmitted independently. “Experimental data indicates that the probes used by Cellmark are independent of one another.”
{State
v.
Pennington, supra,
Thus, the prosecution showed that the method used by Cellmark in this case to arrive at its data base and statistical probabilities was generally accepted in the scientific community. Any question or criticism of the size of the data base or the ratio pertains to weight of the evidence and not to its admissibility.
2. No Due Process Violation
Appellant contends that even though the California Supreme Court has rejected a due process right to retest evidence when retesting is impossible due to the small amount of evidence in the context of blood protein tests (see
People
v.
Griffin
(1988)
3. Sufficiency of the Evidence
Appellant asserts that even if DNA evidence is admitted and given full weight, there was insufficient evidence to support her conviction. We do not look to determine whether appellant’s guilt was established beyond a reasonable doubt, but whether substantial evidence supports the trier of fact’s conclusion in that regard.
(People
v.
Redmond
(1969)
Essentially, appellant asks this court to substitute its judgment for that of the trial court whether to believe witnesses’ prior inconsistent statements concerning appellant’s admissions. The trial court was entitled to find the earlier statements credible.
(People
v.
Thornton
(1974)
Appellant made inconsistent statements to the police about whether she had ever been in the Top Hat and had a motive to commit the crime since she was a cocaine addict and in need of money. Moreover, there was strong circumstantial evidence aside from the DNA test results. Hair identified by the criminalist as similar to hers was found in different places in the Top Hat, including in the victim’s blood and on his clothing. Appellant was seen driving away from the area at the approximate time of the killing. Additionally, her admissions made to or overheard by her sister, father, and sister’s friend added to the equation. With the introduction of the DNA test results, substantial evidence supports the conviction.
That the only witness to have seen someone inside the Top Hat failed to identify appellant does not render the rest of the evidence insufficient.
*870 The judgment is affirmed.
Gilbert, J., and Yegan, J„ concurred.
Appellant’s petition for review by the Supreme Court was denied January 30, 1992. Panelli, J., was of the opinion that the petition should be granted.
*871 Appendix
[[Image here]]
Notes
Red blood cells contain no nuclei and, therefore, no DNA.
(People
v.
Castro, supra,
A chromosome is the location of hereditary (genetic) material within the cell. A gene is the portion of a DNA molecule that comprises the basic, functional hereditary unit
A good example of the product rule is found in
People
v.
Wesley, supra,
State
v.
Pennell, supra,
“Genotype” is the total of the genetic information contained in the chromosomes of an organism; the total genetic makeup of an organism.
We note that the courts in Andrews v. State and Spencer v. Com, stated that the evidence would have been admissible if Frye applied.
We note that since appellant’s brief was filed, the Iowa Supreme Court found that the trial court properly admitted scientific evidence of DNA “fingerprinting” using the reliability standard and also found that the testimony of the state’s expert as to mathematical probability was proper, despite lopsided odds.
(State
v.
Brown
(Iowa 1991)
Courts that have found forensic DNA typing using RFLP analysis generally accepted in the scientific community include
Cobey
v.
State, supra,
Even the defense witnesses here have stated that there is nothing controversial about the theory underlying DNA typing. “Indeed, this theory is so well accepted that its accuracy is *857 unlikely even to be raised as an issue in hearings on the admissibility of the new tests.” (Thompson & Ford, DNA Typing: Acceptance and Weight of the New Genetic Identification Tests (1989) 75 Va. L. Rev. 45, 60.)
Electrophoresis, an integral step in DNA typing, has been established as generally accepted in the scientific community. (See
People
v.
Reilly, supra,
Three commercial laboratories in the United States, Cellmark, Lifecodes, and Cetus, as well as the FBI, currently perform DNA analysis. Both Cellmark and Lifecodes employ RFLP analysis.
Doctor Ford acknowledged that Doctor Roberts was “not just the first name” but “just about the only name that comes to mind in that area.” (Restriction enzyme digestion.)
In
People
v.
Castro, supra,
The probe does not detect a particular allele, but only a repeat sequence.
Cellmark uses this method of measurement rather than standard deviations to account for different thickness in the gel which can cause bands the same distance apart at one end of the gel to appear closer or farther apart than bands the same distance apart at the other end. A fixed standard deviation, i.e., a fixed percentage of fragment size, would not reflect variant band distinguishability along the gel. Doctor Conneally testified that resolution limits are commonly used in biology.
See
State
v.
Pennell, supra,