discussing the science underlying forensic DNA analysisSummary of this case from Commonwealth v. Pruitt
Submitted March 9, 1998.
Decided June 16, 1998.
Appeal from the Court of Common Pleas, Westmoreland County, Criminal Division, No. 638 C 1994, Richard E. McCormick, Jr., J.
Timothy Paul Dawson, Adamsburg, for Donald J. Blasioli.
John W. Peck, Wayne B. Gongaware, Greensburg, for the Com.
Before FLAHERTY, C.J., and ZAPPALA, CAPPY, CASTILLE, NIGRO, NEWMAN and SAYLOR, JJ.
We allowed appeal to determine whether evidence of statistical probabilities calculated using the product rule is admissible at trial in a criminal case to assist the trier of fact in assessing the probative significance of a deoxyribonucleic acid ("DNA") match. We agree with the trial court and the Superior Court that the product rule, as applied in DNA forensic analysis, is generally accepted in the relevant scientific communities and that such evidence therefore meets the standard for admissibility.
In May, 1993, J.D. was assaulted and raped. The crimes occurred late at night, after she departed a neighborhood tavern in Greensburg, Pennsylvania, and while she was walking home to the city of Jeanette along a poorly-lit road. During the attack, the assailant held his hand over J.D.'s eyes, and J.D. closed her eyes throughout the encounter for fear that her assailant would take her life if she saw his face. The assailant smoked a cigarette before departing the scene.
J.D. contacted the police and was taken to the hospital, where medical professionals conducted a rape examination and collected seminal fluid. The Pennsylvania State Police recovered various items from the scene of the attack, including a fresh Bel-Aire cigarette butt, which was tested and found to have been smoked by an individual having type A blood.
In September, 1993, an investigating officer visited Appellant Donald J. Blasioli ("Blasioli") at his home and informed him that an investigation was in progress related to a separate sexual assault that had occurred in the previous month. Upon request, Blasioli provided a saliva sample, but declined to provide hair and blood samples. During the course of the interview, Blasioli admitted that he smoked Bel-Aire cigarettes.
Subsequent testing of Blasioli's saliva sample indicated that he had type A blood. Based upon this information, the police obtained a warrant authorizing them to obtain samples of Blasioli's hair and blood. DNA testing performed at the Pennsylvania State Police laboratory resulted in a determination of a match between Blasioli's blood sample and the semen sample obtained from J.D. immediately after the crimes. Based upon this evidence, Blasioli was arrested and charged with rape, indecent assault, simple assault and unlawful restraint.
Prior to trial, the Commonwealth disclosed its intent to present testimony concerning both the results of the DNA testing and certain probabilities derived from those tests using statistical methods known as the product rule and the ceiling principle. Specifically, the Commonwealth sought to introduce expert testimony that: the probability of a random occurrence in the general population of a DNA profile matching both Blasioli's and the crime sample, calculated using the product rule, was one in 10 billion; and the probability calculated using the ceiling principle was one in 30 million. After a pre-trial hearing pursuant to Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), the trial court ruled that the evidence met the standard of admissibility for novel scientific evidence.
At trial, J.D. testified to the circumstances of the crimes, but was unable to identify Blasioli as the perpetrator. The Commonwealth presented its scientific evidence through the testimony of expert witnesses, and Blasioli presented an expert to refute the Commonwealth's evidence. The jury found Blasioli guilty of all charges, and he was sentenced to concurrent terms of four to eight years imprisonment on the rape charge and six to twelve months on the remaining charges. On direct appeal, the Superior Court affirmed, see Commonwealth v. Blasioli, 454 Pa. Super. 207, 685 A.2d 151 (1996), and this appeal followed.
In determining whether novel scientific evidence is admissible in criminal trials, Pennsylvania courts apply the test set forth in Frye, 293 F. at 1013. See Commonwealth v. Topa, 471 Pa. 223, 231, 369 A.2d 1277, 1281 (1977) (adopting the Frye test). Pursuant to Frye, to be admissible, such evidence must have gained general acceptance in the relevant scientific community. See Commonwealth v. Zook, 532 Pa. 79, 98-99, 615 A.2d 1, 12-13 (1992) (citing Commonwealth. v. Topa 471 Pa. 223, 230, 369 A.2d 1277, 1281 (1977)), cert. denied, 493 U.S. 873, 110 S.Ct. 203, 107 L.Ed.2d 156 (1993). This Court has generally required that both the theory and technique underlying novel scientific evidence must be generally accepted. See generally Crews, 536 Pa. at 522, 640 A.2d at 402 (finding general acceptance with respect to theories and methods of DNA forensic analysis).
As a matter of federal jurisprudence, Frye was overruled in Daubert v. Merrell Dow, 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), on the ground that it had been superceded by the Federal Rules of Evidence. Daubert establishes a two-prong test to determine admissibility of scientific evidence: 1) will the testimony assist the trier of fact; and 2) will the testimony be reliable or scientifically valid? Daubert, 509 U.S. at 592, 113 S.Ct. at 2796. Nevertheless, Pennsylvania courts are not bound by the Federal Rules of Evidence, and, for the present, this Court has continued to employ the Frye standard for determining the admissibility of novel scientific evidence. See, e.g., Commonwealth v. Crews, 536 Pa. 508, 518 n. 2, 640 A.2d 395, 400 n. 2 (1994); see also Pa.R.E. 702, Comment (adopted May 8, 1998; effective October 1, 1998).
In this case, Blasioli attacks the validity of the application of a principle of statistical probability to DNA forensic analysis. Blasioli and the Commonwealth are in apparent agreement that the scientific communities relevant to this issue include the disciplines of population genetics, human genetics and population demographics.
See generally JANE CAMPBELL MORIARTY, 2 PSYCHOLOGICAL AND SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS § F11:23, at 11-22 (1997) [hereinafter MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS]; George Bundy Smith Janet A. Gordon, The Admission of DNA Evidence in State and Federal Courts, 2465, 2467 (May, 1997) [hereinafter Smith Gordon, Admission of DNA Evidence]; see also David H. Kaye, DNA Evidence: Probability, Population Genetics, and the courts, 7 HARV. J. LAW TECH. 101, 101-02 (Fall, 1993) [hereinafter Kaye, DNA Evidence]; Lindsey v. People, 892 P.2d 281, 288-89 (Colo. 1995). The following treatises are cited as authoritative in judicial decisions and secondary sources: MICHAEL J. PELCZAR, JR., ET AL., MICROBIOLOGY: CONCEPTS AND APPLICATIONS (1993) [hereinafter PELCZAR, MICROBIOLOGY]; LANSING M. PRESCOTT ET Al., MICROBIOLOGY (2d ed. 1993) [hereinafter PRESCOTT, MICROBIOLOGY]; R. FRUEND W. WILSON, STATISTICAL METHODS (1993) [hereinafter FRUEND WILSON, STATISTICAL METHODS]; D. STIRZAKER, ELEMENTARY PROBABILITY (1994) [hereinafter STIRZAKER, ELEMENTARY PROBABILITY]; LORNE T. KIRBY, DNA FINGERPRINTING, AN INTRODUCTION (1990) [hereinafter KIRBY, DNA FINGERPRINTING]; R. JOHNSON, ELEMENTARY STATISTICS (4th ed. 1984) [hereinafter JOHNSON, ELEMENTARY STATISTICS]; H. ELDON SUTTON, AN INTRODUCTION TO HUMAN GENETICS (1980) [hereinafter SUTTON, INTRODUCTION TO GENETICS]; and G. BEADLE, THE LANGUAGE OF LIFE (1966) [hereinafter BEADLE, LANGUAGE OF LIFE].
In addressing the merits of the parties' arguments, a brief description of the scientific principles and procedures applied DNA analysis is necessary. DNA is genetic material found most types of cells of the human body, including white blood cells and cells contained in semen and hair follicles. DNA constitutes the primary element of an organism's total genetic information, known as its genome. In the process of cellular division, DNA functions essentially as a template, providing a blueprint for resulting cells. DNA also directs the construction of specific proteins that comprise the structural component of cells and tissues, as well as the production of enzymes necessary for essential biochemical reactions. As such, DNA determines an organism's unique physical composition. See generally MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § 11.11, at 11-8-10; SUTTON, INTRODUCTION TO GENETICS, supra note 2, at 29-60; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2467 (citing PELCZAR, MICROBIOLOGY, supra note 2, at 350-400 (explaining the structure and characteristics of DNA)); Kaye, DNA Evidence, supra note 2, at 107 (citations omitted).
The National Research Council (the "NRC") has generated several primary sources cited almost universally in judicial decisions assessing DNA forensic analysis and the associated statistics. The NRC is a private, non-profit society of distinguished scholars that is administered by the National Academy of Sciences, the National Academy of Engineering and the Institute of Medicine. The NRC formed the Committee on DNA Technology in Forensic Science to study the use of DNA analysis for forensic purposes, resulting in the issuance of a report in 1992. See COMMITTEE ON DNA TECHNOLOGY IN FORENSIC SCIENCE, NATIONAL RESEARCH COUNCIL, DNA TECHNOLOGY IN FORENSIC SCIENCE (1992) [hereinafter 1992 NRC REPORT]; see generally State v. Marcus, 294 N.J. Super. 267, 683 A.2d 221, 227 n. 6 (1996). A new committee was subsequently formed to study recent developments in the field, which also issued a frequently cited report. See NATIONAL RESEARCH COUNCIL, THE EVALUATION OF FORENSIC DNA EVIDENCE 63 (1996) [hereinafter 1996 NRC REPORT]; see generally Marcus, 683 A.2d at 227 n. 6; R. Stephen Kramer, Comment, Admissibility of DNA Statistical Data: A Proliferation of Misconceptions, 30 CAL. W.L.REV. 145, 147 n. 17 (Fall, 1993) (noting that courts have traditionally deferred to pronouncements from the National Academy of Sciences) (citing Rorie Sherman, DNA Unraveling, NAT'L L.J. 1, 30 (Feb. 1, 1993)).
In every nucleated cell in the human body, long strands of DNA are compressed and entwined into bodies called chromosomes, of which there are twenty-three pairs, one-half of each pair in an individual being donated by one's father and the other by the mother. Smith Gordon, Admission of DNA Evidence, supra note 2, at 2466 (citing 1996 NRC REPORT, supra note 3, at 60-63).
A DNA molecule consists of strands in the shape of a ladder, twisted into a characteristic shape resembling a spiral staircase, which is described as a double helix. Each side of the ladder is composed of repeated sequences of phosphate and sugar molecules, with a nitrogen-containing chemical called a base extending toward the opposite side to join a complimentary base, together forming a rung of the ladder. Each unit of a strand, containing one sugar molecule, one phosphate molecule and one base, is called a nucleotide. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § 11.11, at 11-9; Smith Gordon, Admission of DNA Evidence, supra a note 2, at 2466-66 (citing PRESCOTT, MICROBIOLOGY, supra note 2, at 193); Kaye, DNA Evidence, supra note 2, at 107 (citations omitted); see generally Armstead v. State, 342 Md. 38, 673 A.2d 221, 227 (1996) (citing BEADLE, LANGUAGE OF LIFE, supra note 2, at 193-94).
There are four kinds of nucleotide bases in DNA: adenine (A), guanine (G), cytosine (C) and thymine (T). Due to their chemical composition, these can fit together only as follows: adenine will pair only with thymine, and cytosine will pair only with guanine. MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § 11.11, at 11-9; Kaye, DNA Evidence, supra note 2, at 107 n. 33 (citations omitted); see also Smith Gordon, Admission of DNA Evidence, supra note 2, at 2465-66 (citing PRESCOTT, MICROBIOLOGY, supra note 2, at 193). This strict pairing requires that the order of bases on one side of a DNA ladder will determine the order on the other side, establishing the basis for accurate cell reproduction upon splitting of a DNA molecule. Id.
Each pair of nucleotide bases joined to form the rungs of the DNA ladder is called a base pair, of which there are over three billion in any single DNA molecule. It is the unique, repeating sequences of the base pairs along the double strands of DNA that provides the instructions for individual human characteristics. A gene, the fundamental unit of heredity, is a functional unit of DNA containing the base pair sequence responsible for a particular characteristic. The human genome is estimated to comprise at least 100,000 genes. Alternative forms of genes are known as alleles, and the position of a gene or allele on a chromosome is known as its locus. See generally MORIARTY SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS supra note 2, § 11.11, at 11-9; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2466 (citing 1996 NRC REPORT, supra, note 3, at 13).
For example, the gene for the production of eyes may lake the form of a blue-eyed allele or a green-eyed allele. The difference between the alleles results from the sequence of the base pairs along the DNA strands. Smith Gordon, Admission of DNA Evidence, supra note 2, at 2466 (citing 1996 NRC REPORT, supra note 3, at 13-14) Each parent contributes one copy of each gene, so every individual has two copies or alleles of each gene. Id.
Large segments of human DNA are the same from person to person, accounting for human characteristics that are generally shared. Indeed, from the sequence of the 3 billion base pairs, only about 3 million differ from one individual to another (except in the case of identical twins, who have identical DNA). See Smith Gordon, Admission of DNA Evidence, supra note 2, at 2466 (citing 1996 NRC REPORT, supra note 3, at 63). It is the existence of such differences in the sequencing of base pairs, known as "polymorphisms," that provides the basis for DNA identification. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS supra note 2, § 11.11, at 11-9; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2467 (citing 1996 NRC REPORT, supra note 3, at 61); see also Armstead, 673 A.2d at 227-28 (citing same).
Such identification is also referred to as DNA identity testing, profiling, fingerprinting, typing or genotyping. See Kramer, Admissibility of DNA, supra note 3, at 145 n. 1 (citing KIRBY, DNA FINGERPRINTING, supra note 2, at 1).
The length of each polymorphism is determined by the number times a particular base pair sequence is repeated along the chromosome. Stretches of DNA along which a short nucleotide sequence is repeated are known as "variable number tandem repeats" or "VNTRs." Because of their length, such discrete portions of a DNA sample's patterned chemical structure are most easily capable of identification, and much of DNA forensic analysis relies upon loci containing these polymorphisms. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § 11.11, at 11-9-10; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2467 (citing 1992 NRC REPORT, supra note 3, at 61); Kaye, DNA Evidence, supra note 2, at 108-09 (citations omitted).
DNA forensic analysis begins with the preparation of a DNA profile, which entails the creation of a picture of multiple VNTRs. One of several techniques is used, among which is the restriction fragment length polymorphism method (the "RFLP method"), which was used by the State Police laboratory in this case and which is commonly used by the FBI and law enforcement laboratories across the country. The method isolates VNTRs known as restriction fragments by the use of restriction enzymes, chemical "scissors" that recognize short base pair sequences and cut DNA molecules at those specific sites. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.14, at 11-11; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2469 (citing PRESCOTT, MICROBIOLOGY, supra note 2, at 288-89); see also KIRBY, DNA FINGERPRINTING, supra note 2, at 55. Once the restriction fragments are chemically sorted according to size, x-ray pictures are created known as autorads, using the process of autoradiography. The autorad displays a discernible pattern of dark bands resembling an electronic bar code, each band representing a fragment of DNA. Id.
This is accomplished by a process called "agarose gel electrophoresis," which involves passing a current through a gel medium containing the fragments. The negatively-charged RFLPs migrate toward a positive electrode. Because their progress through the gel is dependent upon their size, the fragments separate according to their length. See generally MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.15, at 11-11-12; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2469-70 (citing PRESCOTT, MICROBIOLOGY, supra note 2, at 288-89); Kaye, DNA Evidence, supra note 2, at 108 (citations omitted); Blasioli, 454 Pa. Super. at 229-30, 685 A.2d at 162-63 (elaborating upon the details of the RFPL method).
Profiling also entails transferring the restriction fragments to a nylon membrane in the same pattern as in the gel (known as Southern Blotting), chemically separating DNA strands (denaturing), addition of one or more radioactive markers or probes to bind with the single DNA strands to aid in identification (hybridization), followed by the autoradiography. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.16-18, at 11-12-13; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2469-70 (citing 1996 NRC REPORT, supra note 3, at 38-39); see also KIRBY, DNA FINGERPRINTING, supra note 2, at 26, 94-101; Armstead, 673 A.2d at 225-26 (citations omitted): Lindsey, 892 P.2d at 286 (citing William C. Thompson Simon Ford, DNA Typing, 75 VA.L. REV. 45, 74-76 (1989)); see generally 1992 NRC REPORT, supra note 3, at 3-5.
A newer method of DNA analysis has been developed, known as the polymerase chain reaction method ("PCR"), which uses the same process by which cells replicate to amplify a small quantity of DNA for purposes of analysis. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11. 19, at 11-13-14; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2470-71 (citations omitted); see also Armstead, 673 A.2d at 228 n. 9 (citing 1992 NRC REPORT, supra note 3, at 1-8 to 1-10). That method apparently is gaining acceptance in the scientific community, see id.; however, it was not at issue in either Crews or in this case.
After DNA profiles are created for both the crime scene and suspect samples, the autorad patterns are measured and compared according to their length. If the similarities are such that they fall within a narrow margin, known as a match window, the samples are declared a match. See MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.20, at 11-14; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2470, 2473 (citing Kaye, DNA Evidence, supra note 2, at 110-11).
According to the evidence in this case, the Pennsylvania State Police laboratory defines the range which constitutes a match as plus or minus 2.5 percent of the sample length.
The evidence in this ease established that the Pennsylvania State Police laboratory focuses upon VNTRs from six different polymorphic sites to determine whether an overall match exists. Examination of DNA fragments from multiple loci reduces the probability of a random occurrence of the overall profile (a "random match"). See Smith Gordon, Admission of DNA Evidence, supra note 2, at 2473 (citing 1996 NRC REPORT, supra note 3, at 4-5); see generally B.S. Weir S.S. Gaut, Matching and Binning DNA Fragments in Forensic Science, 34 JURIMETRICS 9, 10 (1993).
The general acceptance in the discipline of human genetics of DNA analysis using the RFLP method was recognized in Crews, 536 Pa. at 519-20, 640 A.2d at 400-401, and has been clearly shown by authoritative scientific literature, the overwhelming weight of judicial authority and by the testimony of the Commonwealth's experts at the Frye hearing in this case. See generally Marcus, 683 A.2d at 229 (citing cases); Thomas J. Fleming, Annotation, Admissibility of DNA Identification Evidence, 84 A.L.R.4th 313 § 4 (1991 Supp. 1997). Crews, however, did not answer the question of the admissibility of associated statistical evidence, and controversy has existed within relevant scientific communities as to the validity of the application of such statistical methods to DNA forensic analysis.
In Crews, after concluding that evidence of physical DNA forensic analysis was admissible, this Court observed that:
[w]hat has not yet achieved universal agreement is the less objective selection of the appropriate population for statistical purposes and the actual statistical analysis which is to be applied to the physical analysis carried out in the laboratory. About the statistical treatment of the physical evidence there remains disagreement and continuing theoretical development. Crews, 536 Pa. at 520, 640 A.2d at 401.
The statistical assessment performed after a match has been declared is called population frequency analysis. The object is to determine the overall likelihood that someone other than the suspect would possess DNA matching that in the sample obtained from the crime scene. The first step is to determine, for each matching allele, the likelihood that such an allele would appear in a randomly selected individual. See Smith Gordon, Admission of DNA Evidence, supra note 2, at 2472 (citing 1996 NRC REPORT, supra note 3, at 74); Lindsey, 892 P.2d at 286. This determination is made through the application of theoretical models based upon population genetics. Id.
The analysis does not yield the probability that a particular defendant is the source of the crime sample or committed the crime at issue; instead, it provides only an estimate of the probability that a randomly selected person from the general population would genetically match the crime sample as well as the suspect. See generally Smith Gordon, Admission of DNA Evidence, supra note 2, at 2472 (citing Jonathan J. Keohler, DNA Matches and Statistics: Important Questions, Surprising Answers, 76 JUDICATURE 222 n. 1, 224 (1993)); see also Lindsey, 892 P.2d at 290 (citing Jonathan J. Keohler, Error and Exaggeration in the Presentation of DNA Evidence at Trial, 34 JURIMETRICS J. 21, 31-32 (1993)).
Such models are generated by creation of a computer database containing DNA profiles obtained from the general population. The frequency of an allele obtained from a sample can be determined by calculating the probability that a matching allele would appear in a DNA sample obtained from an individual who was randomly selected from the database. See MORIARTY SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.23, at 11-16; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2474 (citations omitted); see generally Armstead, 673 A.2d at 235 (citing 1992 NRC REPORT, supra note 3, at 2-3).
To ameliorate theoretical problems associated with population substructures, discussed below, the Pennsylvania State Police laboratory database categorizes DNA samples according to three racial groups, and uses a process known as "fixed binning." The probability of random matching is also reduced by choosing highly variable segments of the DNA, with dozens of individual alleles, so that individual allele frequency will be very low. Additional variations occur in the matching of the maternal and paternal alleles located at each locus, further reducing the probability of a random match. See generally Armstead, 673 A.2d at 236 (citing R. Chakraborty K. Kidd, The Utility of DNA Typing in Forensic Work, 254 SCIENCE 1735, 1735 (1991)[hereinafter Chakraborty Kidd, Utility of DNA Typing]).
The classifications are: Caucasian, African-American and Hispanic. The database has approximately 1140 samples (500 Caucasian, 330 African-American and 310 Hispanic) and was compiled mostly from the three major metropolitan areas of the state.
Binning is the process of grouping similarly sized alleles within a database and assigning population frequencies to each such group. Upon measurement of an allele from a crime sample, it is assigned a bin and the corresponding frequency assigned to that bin. If the allele falls between bins, it is assigned to the bin with the greater frequency, resulting in the more conservative probability estimate. Binning is used to account for variables in recording autorads and to provide confidence limits on frequency estimates. See Kaye, DNA Evidence, supra note 2, at 122-23 (citations omitted); Lindsey, 892 P.2d at 287 n. 18; Armstead, 673 A.2d at 236 n. 21 (citing L. Mueller, Population Genetics of Hypervariable Human DNA, in FORENSIC DNA TECHNOLOGY 56 (1992)). It also has the advantage of permitting individual calculation by reference to a table, rather than by making independent calculations in each case. See generally Kaye, DNA Evidence, supra note 2, at 124 (citations omitted); Blasioli, 454 Pa. Super. at 233, 685 A.2d at 164.
See infra note 17. Blasioli's challenge in this case focuses upon application of the product rule, not upon the use of the Pennsylvania State Police laboratory database to calculate individual allele frequency. Nevertheless, we note substantial agreement in the scientific community that databases on the order of that maintained by the Pennsylvania State Police laboratory are adequate for estimating allele frequencies. See generally State v. Copeland, 130 Wn.2d 244, 922 P.2d 1304, 1320-22 (1996) (citing scientific journals). Several courts have therefore appropriately concluded that questions about the size of a database, as well as other issues which may arise in connection with the DNA analysis of particular evidence samples, generally go to the weight of the evidence, rather than to its admissibility. Id. at 1321; see also Marcus, 683 A.2d at 227 n. 7; Lindsey, 892 P.2d at 292-93.
Once the probability of random occurrence is calculated for each individual allele, the individual probabilities may be combined to determine an overall probability of random matching across the genetic profile. In order to make this calculation, scientists have employed the product rule. The product rule states that the probability of two events occurring together is equal to the probability that the first event will occur multiplied by the probability that the second event will occur. See Kaye, DNA Evidence, supra note 2, at 127-28 (citations omitted); Armstead, 673 A.2d at 236 (citing FRUEND WILSON, STATISTICAL METHODS, supra note 2, at 62). Coin tossing is commonly used as an illustration — the probability of a coin flip resulting in "heads" on successive tries is equal to the probability of the first toss yielding heads, fifty percent, times the probability of heads on the second toss, fifty percent, equaling twenty-five percent. See Armstead, 673 A.2d at 236 (citing JOHNSON, ELEMENTARY STATISTICS, supra note 2, at 143).
Stated as such, the product rule is not the type of novel scientific evidence to which Pennsylvania courts apply the Frye test. Application in the context of DNA forensic analysis, however, as further discussed below, represents a use that combines statistical principles with novel forensic analysis and entails the "type of manipulation of physical evidence that requires evaluation under Frye." Lindsey, 892 P.2d at 290.
As applied in DNA typing, the product rule states that the probability of a genetic profile occurring randomly is the product of the probabilities of each individual allele's occurrence in the general population. MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.24, at 11-17; see also 1992 NRC REPORT, supra note 3, at 76. Such application can produce odds of up to one in 739 billion of a random profile match. Kramer, supra, at 146 n. 3 (citing Ira Pilchen, Federal Report and Court Rulings Intensify DNA Evidence Debate, JUDICATURE 41 (June 1992)).
Frequency calculation is also complicated by the fact that two alleles, maternal and paternal, are located at any particular DNA locus. Prior to the general application of the product rule, a calculation is performed to determine the frequency for the pair of alleles (maternal and paternal) located at each locus. The calculation is different, depending upon whether the two alleles match (or are homozygous), or do not match and are thus said to be heterozygous. See generally Lindsey, 892 P.2d at 287; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2466-67; Kaye, DNA Evidence, supra note 2, at 124.
Valid use of the product rule in any context depends upon the statistical independence of each component factor of the equation. Smith Gordon, Admission of DNA Evidence, supra note 2, at 2475 (citing Kaye, DNA Evidence, supra note 2, at 122 n. 93); see also Armstead, 673 A.2d at 236-37 n. 22 (citing STIRZAKER, ELEMENTARY PROBABILITY, supra note 2, at 22-30 (discussing independent versus dependent events)). Independence exists where the outcome of the first event does not impact upon the outcome of the second event. Id. Validity of the rule in DNA forensic analysis depends upon whether individual alleles are actually statistically independent, requiring that the probability of finding one allele is not significantly affected by having found any other allele. Smith Gordon, Admission of DNA Evidence, supra note 2, at 2475 (citing Kaye, DNA Evidence, supra note 2, at 122 n. 93); see also 1992 NRC REPORT, supra note 3, at 77.
The NRC illustrated this independence principle by referring to the example of individuals in the Nordic population with alleles for blonde hair, blue eyes and fair skin. If each trait carried a one-in-ten probability of occurrence, application of the product rule would result in a one-in-one-thousand probability that an individual would have all three traits. The combination, however, occurs commonly in the Nordic population, raising concerns about the validity of the application of the product rule to allele frequency calculations. See 1992 NRC REPORT, supra note 3, at 76-77.
In this case, Blasioli maintains that use of the product rule is not generally accepted in relevant scientific communities, in particular because of an asserted lack of statistical independence of allele frequency. For many years, some scientists argued that the product rule can validly be applied only where members of racial and ethnic groups represented by a database intermix randomly, without regard to religion, ethnicity or geography. The view was premised upon the theory that population substructures affect the frequency of alleles and undermine the independence of such genetic factors and, hence, valid application of the product rule. See generally R. Lewontin D. Hartl, Population Genetics in Forensic DNA Typing, 254 SCIENCE 1745, 1745-46 (December 20, 1991) [hereinafter Lewontin Hartl, Population Genetics]; see also Smith Gordon, Admission of DNA Evidence, supra note 2, at 2475 (citing 1996 NRC REPORT, supra note 3, at 77); Kaye, DNA Evidence, supra note 2, at 128 (citations omitted); Marcus, 683 A.2d at 228 (citing cases); Lindsey, 892 P.2d at 288; People v. Miller, 173 Ill.2d 167, 219 Ill.Dec. 43, 670 N.E.2d 721, 731-32 (1996). Among these critics were two prominent population geneticists, Professors Daniel Hartl and Richard Lewontin. See Lewontin Hartl, Population Genetics, supra, at 1745-46. Furthermore, in its 1992 report, the NRC noted the existing dispute and proposed that a conservative modification of the product rule, known as the ceiling principle, be used in calculating the probability of a genetic match. See 1992 NRC REPORT, supra note 3, at 13.
Population substructures occur when a certain subsection of the population, for example, a racial or ethnic group, selectively mates within that same subsection, resulting in the interrelation of certain genetic traits. Several theories developed which suggested that truly random mating across racial and ethnic lines was necessary to independence in the distribution of individual alleles in a population. The terms "linkage equilibrium," "gametic phase balance," and "Hardy-Weinberg equilibrium" identify versions of such theories. See generally MORIARTY, SCIENTIFIC EVIDENCE IN CRIMINAL TRIALS, supra note 2, § F11.26-28, at 11-18-20; Smith Gordon, Admission of DNA Evidence, supra note 2, at 2476-77 (citations omitted); see also Armstead, 673 A.2d at 237 n. 23 (citing Lewontin Hartl, Population Genetics, supra, at 1746-47). For example, Hardy-Weinberg equilibrium referred to a state in which one allele at a locus is not predictive of the other allele at that locus (one allele is inherited from the mother, the other from the father). Smith Gordon, Admission of DNA Evidence, supra note 2, at 2475 (citing 1992 NRC REPORT, supra note 3, at 77-79); Copeland, 922 P.2d at 1317 (citations omitted); see also Lindsey, 892 P.2d at 287 n. 21 (citations omitted).
The ceiling principle assumes the existence of some degree of population substructuring and generates more conservative population probability statistics than the product rule by using the maximum frequency of allele occurrences from random samples from racial and ethnic subgroups. Marcus, 683 A.2d at 228 (citing 1992 NRC REPORT, supra note 3, at 13); see also Smith Gordon, Admission of DNA Evidence, supra note 2, at 2476 (citing Kaye, DNA Evidence, supra note 2, at 134); 1992 NRC REPORT, supra note 3, at 82-84. The largest of these frequencies, or ten percent, whichever is greater, is then used to compute the probability of a random match (the 1992 NRC report recommended using the ten percent figure as an interim measure, to be replaced by a figure of five percent once preliminary data was collected). According to the NRC, the ceiling rates were considerably higher than the likely maximum allele frequency for any subgroup. See 1992 NRC REPORT, supra note 3, at 93. The resulting probabilities for each allele occurrence are then multiplied as in product rule computations. Id.
Importantly, the NRC's 1992 report did not constitute an outright rejection of the product rule. Instead, the NRC merely recommended that, until data could be assembled from which to assess the impact of any significant population substructuring, the ceiling principle could be applied to impose an appropriate degree of conservatism. See Armstead, 673 A.2d at 237 (citing 1992 NRC REPORT, supra note 3, at S-10 to S-11). The suggestion in the 1992 NRC Report, however, that substructuring could impact upon the validity of the product rule in DNA forensic analysis persuaded a number of courts that general agreement in the scientific community was lacking. See, e.g., Copeland, 922 P.2d at 1317-18 (citing State v. Cauthron, 120 Wn.2d 879, 846 P.2d 502 (1993)). While substantial debate ensued, it is noteworthy that no empirical data existed to support theories postulating a substantial impact of substructuring upon DNA forensic analysis. See Chakraborty Kidd, Utility of DNA Typing, supra, at 1735, cited in Armstead 673 A.2d at 237; see also Kaye, DNA Evidence, supra note 2, at 168 (stating that "[t]here is very little evidence, and certainly no scientific consensus, that the impact [of substructuring] is substantial in any known population").
Several events subsequently occurred, indicating that the controversy over the use of the product rule has dissipated. In 1993, the FBI conducted an extensive, international study of VNTR frequency data. See LABORATORY DIVISION, FEDERAL BUREAU OF INVESTIGATION, UNITED STATES DEPARTMENT OF JUSTICE, I-A VNTR POPULATION DATA: A WORLDWIDE STUDY 2 (Feb. 1993) [hereinafter DEP'T OF JUSTICE, VNTR POPULATION DATA]. The study concluded that population frequency calculation using the product rule was reliable, valid and meaningful, without forensically significant consequences resulting from population substructure as had been postulated by some scientists. Id.; see generally Lindsey, 892 P.2d at 294 (emphasis supplied) (citing DEP'T OF JUSTICE, VNTR POPULATION DATA, supra, at 2); see also Copeland, 922 P.2d at 1319.
It is important to note that the relevant question is not whether some such substructuring exists, but whether the deviations it induces have an appreciable effect upon the relative frequency of the particular, highly-variable alleles selected for DNA profiling. See supra pages 11-12; see generally Kaye, DNA Evidence, supra note 2, at 169 (citations omitted).
Additionally, in 1994, Dr. Eric Lander, a former leading opponent of the use of the product rule, coauthored an article in which he declared that the "DNA fingerprinting wars are over." E. Lander B. Budowle, DNA Fingerprinting Dispute Laid to Rest, 371 NATURE 735, 735 (Oct. 27, 1994) [hereinafter Lander Budowle, DNA Fingerprinting Dispute Laid to Rest]. In the article, the authors stated that the 1992 NRC Report "failed to state clearly enough that the ceiling principle was intended as an ultra-conservative calculation, which did not bar experts from providing their own `best estimates' based on the product rule." Lander Budowle, DNA Fingerprinting Dispute Laid to Rest, supra, at 737, cited in Copeland, 922 P.2d at 1319. They noted that the FBI's laboratory maintained a "remarkable" database, and that, reassuringly, "observed variation is modest for the loci used in forensic analysis and random matches are quite rare, supporting the notion that the FBI's implementation of the product rule is a reasonable best estimate." Id.; see also Lindsey, 892 P.2d at 293.
In their article, the authors emphasized the convergence of scientific opinion concerning population genetics statistics, noting that Budowle was one of the principal creators of the FBI's DNA-typing program and that Lander was an early critic of the lack of scientific standards in DNA-typing and was on the NRC committee, and concluding that "it is fair to say that we represent the range of scientific debate." Lander Budowle, DNA Fingerprinting Dispute Laid to Rest, supra, at 735.
Furthermore, extensive literature in peer-reviewed journals accumulated to support the premise that substructuring does not impact significantly upon DNA population frequency estimates. See Kaye, DNA Evidence, supra note 2, at 126 n. 113, 129-30, 161 (citing scientific journals espousing the view that statistical tests demonstrate the independence of VNTR alleles and arguing that "suitably computed and presented match-binning frequencies and probabilities pass muster under conventional rules of evidence"); Copeland, 922 P.2d at 1319 (citing scientific journals); Armstead, 673 A.2d at 238-39 (same); B. Budowle et al., The Assessment of Frequency Estimates of Hae III-Generated VNTR Profiles in Various Reference Databases, 39 J. FORENSIC SCIENCE 319, 349 (1994) (stating that "[s]ubdivision, either by ethnic group or by U.S. geographic region, within a major population group does not substantially affect forensic estimates of the likelihood of occurrence of a DNA profile").
Additionally, in 1996, the NRC reexamined the methodology issue and also concluded that the use of the ceiling principle for forensic purposes is unnecessary, not only because the principle overstates the effect of population substructuring, but also because of the current abundance of data regarding different ethnic groups within the major races. 1996 NRC REPORT, supra note 3, at 5-30 to 5-35. The 1996 NRC Report reaffirmed the conclusion of the 1992 report that properly conducted DNA tests produce highly reliable results, and that DNA analysis, including the application of statistical probabilities, is generally accepted in relevant scientific communities. Id. at 2-4. Accordingly, "[t]he Committee now recommends the use of a modified version of the product rule which assumes the existence of some undetected population substructure of a lesser magnitude than that reflected by use of the ceiling principle." Marcus, 683 A.2d at 228 (citing 1996 NRC REPORT, supra note 3, at O-21).
A majority of jurisdictions have acknowledged these developments — including the FBI study, the article by Lander and Budowle, and the 1996 NRC report — and have concluded that the controversy over the use of the product rule has been sufficiently resolved. See, e.g., Armstead, 673 A.2d at 221; Copeland, 922 P.2d at 1304; Miller, 219 Ill.Dec. 43, 670 N.E.2d at 721; Marcus, 683 A.2d at 221; State v. Morel, 676 A.2d 1347 (R.I. 1996); Lindsey, 892 P.2d at 281; State v. Dinkins, 319 S.C. 415, 462 S.E.2d 59 (1995); State v. Weeks, 270 Mont. 63, 891 P.2d 477 (1995); State v. Anderson, 118 N.M. 284, 881 P.2d 29 (1994); State v. Futrell, 112 N.C. App. 651, 436 S.E.2d 884 (1993); People v. Chandler, 211 Mich. App. 604, 536 N.W.2d 799 (1995); Taylor v. State, 889 P.2d 319 (Okla.Ct.App. 1995).
We recognize that several of these jurisdictions employ the federal standard enunciated in Daubert, which has been characterized as more lenient than the Frye test.
At the Frye hearing in this case, the Commonwealth presented evidence of general acceptance of the product rule in the relevant scientific disciplines. Such evidence included citation to numerous scientific texts and journals and the testimony of professors of human genetics and statistics from prominent universities. At trial, Blasioli was permitted to contest the Commonwealth's DNA forensic evidence, including the statistical expressions based upon the product rule and the ceiling principle. Blasioli did so through the testimony of an expert who emphasized the theoretical impact of population substructuring along the lines advanced by Lewontin and Hartl. See generally Lewontin Hartl, Population Genetics, supra, at 1745-46. The expert also offered an alternative analysis known as the counting method, whereby he determined that the chances of another genetic match in this case were 1 in 2,220.
Notably, Blasioli did not challenge the admissibility of the statistical evidence calculated using the ceiling principle.
This method is simply an enumeration of how many times an event occurred in a given set of observations.
While we are cognizant of the fact that unanimity among scientists does not exist, unanimity is not required for general acceptance. See Copeland, 922 P.2d at 1319; Lindsey, 892 P.2d at 289. Certainly the relevant science will continue to evolve and techniques will be refined or change. At present, however, it is clear from the scientific commentary, the clear weight of judicial authority, and the evidence in this case that the product rule has gained general acceptance across the disciplines of population genetics, human genetics and population demographics. As such, any remaining dispute as to the validity of the product rule should not result in the exclusion of evidence based upon this statistical method in criminal trials in Pennsylvania.
There is a split among Frye jurisdictions as to whether the test should be applied to determine general agreement in the scientific community as of the time of trial or as of the time of appellate review. Compare Lindsey, 892 P.2d at 291 n. 25, with State v. Bible, 175 Ariz. 549, 858 P.2d 1152, 1189 n. 33 (1993), cert. denied, 511 U.S. 1046, 114 S.Ct. 1578, 128 L.Ed.2d 221 (1994). As noted above, we find the evidence in this case sufficient to establish general acceptance as of the time of trial.
In sum, we hold that statistical evidence based upon the product rule was properly admitted at the trial in this case. Accordingly, the judgment of sentence is affirmed.