JEFF BROWN
Public Defender
MICHAEL BURT
Deputy Public Defender
555 Seventh Street, Second Floor
San Francisco, California 94103
(415) 553-9650
Attorneys for Defendant
ROBERT NAWI
SUPERIOR COURT OF CALIFORNIA
COUNTY OF SAN FRANCISCO
THE PEOPLE OF THE STATE OF ) CALIFORNIA, )
)
Plaintiff, ) )
vs. )
)
ROBERT NAWI )
)
Defendant. )
___________________________________ )SCN: 176527
MCN: 1812436
Date: January__, 2000
Time:
Department:
OPPOSITION TO PEOPLE'S MOTION TO ADMIT DNA TEST RESULTS AND REQUEST
FOR KELLY-FRYE HEARING
TO TERENCE HALLINAN, DISTRICT ATTORNEY FOR THE CITY AND COUNTY
OF SAN FRANCISCO, STATE OF CALIFORNIA, TO ASSISTANT DISTRICT ATTORNEY
JOHN FARRELL AND TO THE ABOVE-ENTITLED COURT:
PLEASE TAKE NOTICE that on ,________ 2000 at a.m. in Department __ of
the above-entitled court, the defendant herein will oppose the People's
anticipated Motion to Admit DNA Test Results and will and hereby does
request a Kelly-Frye hearing on the grounds that (1) the People's DNA
evidence is inadmissible under Kelly-Frye and under Evidence Code
Section 352; (2) this opposition raises issues pursuant to Kelly-Frye
that have not been decided by any California appellate court; and, (3)
the defendant is entitled in any event to a hearing to determine whether
correct scientific DNA testing procedures were used in this particular
murder case.
This Motion is based on this Notice, the attached memorandum of points
and authorities, the testimony and exhibits in the Kelly-Frye hearing in
People v. Bokin et al, S.C. No. 168461, judicial notice of which is
hereby requested pursuant to Evidence Code Section 452(d), and any oral
and documentary evidence and argument as may be produced at the hearing
on said motion.
DATED: December 20, 1999
Respectfully Submitted,
____________________________
MICHAEL N. BURT
Attorneys for Defendant
ROBERT NAWI
I.
INTRODUCTION
When science is imported from the laboratory into the courtroom the leap
from one context to another must be carefully analyzed. Because DNA
evidence based on seemingly astronomical statistical probabilities by
its very nature is particularly compelling to lay jurors, and because
DNA typing is used in the criminal context for purposes of identifying
perpetrators of particularly heinous crimes, extreme circumspection in
evaluating the techniques and protocols of DNA typing is mandated. See
People v. Venegas (1998) 18 Cal.4th 47, 81-84 (noting "DNA evidence
is different", and that "[t]his is an instance in which the
method of scientific proof is so impenetrable that it would ‘...assume
a posture of mystic infallibility in the eyes of a jury...'");
People v. Kelly (1976) 17 Cal.3d 24, 31-32 ("When identification is
chiefly founded upon an opinion which is derived from utilization of an
unproven process or technique, the court must be particularly careful to
scrutinize the general acceptance of the technique.") In addition,
because DNA typing methodologies for use in the forensic arena has been
the subject of several appellate decisions, keen attention to the issues
that have been decided, and those that are still unresolved is required.
A. FUNDAMENTALS OF DNA
DNA is the carrier of genetic information in every plant, animal, and
mircroorganism including bacteria. DNA consists of two strands that are
held together at regular intervals to form a double helix. A useful, and
common, analogy is a ladder: a DNA molecule looks like a twisted ladder
or a spiral staircase. The rungs are made up of two bases that are bound
together, and are attached to strands of sugar molecules. These strands
are the sides of the ladder. There are only four different bases:
adenine (A), guanine (G), cytosine (C), and thymine (T). The base A can
only bind with the base T and C only with G. The binding qualities of
the bases result in two complementary strands (picture the ladder split
from top to bottom) that are not identical. For example, if the sequence
of bases on one strand (one half of the ladder) is ATTGCT, the bases on
the other half of the ladder will have the sequence TAACGA.
When a human being or any other living organism grows, it must duplicate
its DNA so that each new cell will have a complete copy of that
organism's DNA. The first step of the duplication, or reproduction,
process is that the DNA molecule is first "unzipped": the
strand splits lengthwise. The bonds between the two bases are broken by
enzymes and the DNA becomes single-stranded. Enzymes present in the cell
then make a new strand of DNA using the two existing strands as
templates, resulting in two complete double-stranded pieces of DNA.
Genes, often called markers, are the genetic units responsible for
inherited traits such as eye and hair color and genetic diseases such as
sickle cell anemia. Genes consist of specific sequences of bases. In
other words, a gene refers to a particular part of the DNA ladder,
identified by the rungs at that part of the ladder. The base sequence of
a gene is generally unique to that gene. However, a gene can have
variations, usually small changes in the base sequence resulting in a
different type of allele of the same gene or marker. Each
double-stranded segment of DNA contains only one variant or allele.
Genes are organized in the human cell on structures called chromosomes,
which are made up of DNA molecules and found in the nucleus of the cell.
Humans have twenty-three chromosomes and each cell has two copies of
each chromosome, resulting in a total of forty-six chromosomes. In
humans, a person inherits twenty-three chromosomes from their mother and
twenty-three chromosomes from their father. A gene exists at only one
location on a single chromosome. Thus, an individual has two copies or
alleles of each gene, one on either chromosome. Even if the two copies
are the same variant, they are called alleles of the same gene.
For example, consider a fictional gene, Gene Z. This gene is found on
human chromosome six. There are two different variants or alleles of Z
found in the human population, Z-1 and Z-2. If Individual A is
homozygous for Z-1, that means that Individual A has two copies of Z-1,
one on each copy of chromosome six. If Individual B is heterozygous for
Z that means Individual B has allele Z-1 on one copy of chromosome six
and has allele Z-2 on the other copy of chromosome six.
B. DNA TYPING TECHNOLOGY
"DNA typing" is a catch-all term for a wide range of methods
for studying genetic variations. Advances in DNA technology paved the
way for detection of variation (polymorphism) in specific DNA sequences.
DNA typing is the analysis of a DNA sample to identify either variations
in the length of specific pieces of DNA (known as RFLP typing) or
whether or not a specific piece of DNA is present (known as PCR typing).
DNA can be extracted from various bodily fluids, such as semen and
blood, and the amount of DNA obtained is proportional to the number of
nucleated cells present in the sample. RFLP requires a relatively large
sample, and is commonly known as "DNA fingerprinting" as it
can result in a positive exclusion of an individual. PCR requires a
relatively small sample, and also can be used to positively exclude a
donor.
RFLP examines loci characterized by a variable number of long tandem
repeats (VNTRs). At some DNA loci, a core sequence of base pairs,
typically 15 to 35 base pairs in length, is repeated a number of times
consecutively along the chromosome. Alleles differ in number of repeats
and are differentiated on the basis of size. Steps in the RFLP analysis
include restriction enzyme digestion of the DNA; separation of DNA
fragments by electrophoresis, Southern blotting, and probing with
locus-specific radioactive or chemiluminescent probes. The result is a
gel with bands; the bands show the lengths of loci from the samples
taken from known and unknown sources. In other words, the gel allows the
analyst to compare the lengths of the VNTRs of samples taken from the
crime scene, including from the victim's body, with samples taken from
the victim and the defendant. Once the analyst determines that the
length of the VNTRs "match" the analyst must then determine
the significance of that match. Specifically, the analyst determines,
using databases of known samples, what is the likelihood that the
reported match is a true match. The results often are extremely high
numbers, and represent the chance that there is a single randomly
selected person would match the trace evidence. Koehler, supra, at 30.
Forensic DNA typing evolved from medical diagnostic techniques. Medical
diagnostic DNA typing involves adequate, clean samples from known
sources, a comparison of discrete alternatives, and it requires no
knowledge of the distribution of patterns in the general population.
In contrast, forensic DNA typing involves samples that are often
degraded, contaminated, and from multiple, unknown sources. It also
involves matching of couples from a wide range of alternatives present
in the population and thus lacks built-in consistency checks. And,
except in cases where the DNA evidence excludes a suspected donor,
assessing the significance of an apparent match requires statistical
analyses of population frequencies. (NRC 1 at 27-63.)
C. THE BIOLOGICAL BASIS FOR PCR-BASED TESTING
The natural process of DNA replication has been duplicated in PCR
technology. PCR is not a genetic test, "but merely a tool to
increase the amount of genetic material to be tested." Thompson,
supra, at 215. The National Research Council describes PCR as "a
laboratory process for copying a chosen short segment of DNA millions of
times." National Research Council, The Evaluation of Forensic DNA
Evidence 69 (1996) [hereinafter "NRC 2"]. Scientists in the
laboratory isolate DNA, and, using heat instead of enzymes, unzip the
strands of DNA. The scientist then adds a primer, a short piece of DNA
with a sequence known to the scientist, to the unzipped DNA. The primer
"marks" the section on the DNA molecule that will be
amplified. An enzyme called polymerase is added to the primer-DNA
mixture. As in the living cell, the polymerase creates a new
complementary strand from each of the original strands present. This
simple process is repeated multiple times, each time separating and
duplicating the existing DNA. In this way , large quantities of a
specific DNA sequence can be made from a very small amount of sample
DNA.
The quality of the copy created by PCR is only as good as the quality of
the original sample. Contamination of samples by bacteria or other
substances containing DNA affect the reliability of the amplification.
NRC 1 at 64-65; NRC 2 at 82-83. In other words, any contamination in a
sample will be amplified along with the DNA from the sample. "[A]mplification
of contaminant could lead to spurious typing results." NRC 1 at 65.
Again, PCR is used to replicate small segments of the DNA strand.
Scientists have identified several small segments, or loci, where the
DNA strand varies among groups of people. Where a locus is highly
variant, it is called polymorphic.
Simply because a locus is polymorphic does not mean that that locus is
appropriate for use in forensic DNA testing. See NRC 1 at 63 (noting
that "some loci can cause systematically unfaithful
amplification"). Several criteria should be met before a locus is
deemed appropriate for forensic DNA testing. These criteria include:
1. The marker should be highly polymorphic and have a high level of
genetic heterozygosity;
2. The target sequence should be easily and specifically amplified;
3. Methods for detecting allelic variation should be uncomplicated and
thoroughly reliable;
4. Population data on genotype frequencies must be available in order to
assign estimates of the marker's power of discrimination and the
probability of false inclusion;
5. The marker systems should be inherited independently so that
frequencies derived from on marker system can be multiplied with those
from others, thereby increasing the power of discrimination. Independent
inheritance occurs when the markers are on separate chromosomes or are
in linkage inequilibrium when present on the same chromosome.
Kamrin T. MacKnight, Comment, The Polymerase Chain Reaction (PCR): The
Second Generation of DNA Analysis Methods Take the Stand, 9 Santa Clara
Computer and High Tech. L.J. 287, 307 (1993) citing Cecilia H. von
Beroldingen et al., Applications of PCR to the Analysis of Biological
Evidence, in PCR Technology: Principles and Applications for DNA
Amplification 209, 201 (Henry A. Erlich ed. 1989).
Loci used in forensic testing are of two types: those loci that vary by
sequence and those that vary by length. Sequence variants are
differences in the actual sequence of bases at a particular location on
the DNA molecule. Length variants look at the number of times a sequence
of bases repeats. NRC 2 at 74. The San Francisco Crime Lab conducts
forensic DNA typing on both sequence variants and length variants.
Tests used to identify sequence variants in a DNA sample employ a
reverse-dot blot method. Probes (single stranded pieces of DNA that are
complementary to the genes or markers amplified) are fixed to a membrane
in dots along a strip. The probes are "synthetic" copies of
the alleles of each gene or marker used. The amplified DNA which has
been denatured is applied to the strips in a buffered solution and the
amplified DNA is allowed to bind to its complementary probe on the
strip. When amplified DNA binds to a particular dot, the dot appears
blue because of a chemical tag that is added to the DNA during the
amplification process. The intensity of the blue color is important in
determining the genotype of the sample. Dot blots appear with varying
intensities, and judging whether the appearance of a faint dot
constitutes the presence of an allele, or simply contamination is an
extremely subjective determination. On the strip, each of the dots is
labeled to correspond to an allele of the marker, and a pattern or
combination of these dots gives the genotype of the person.
HLADQ [hereinafter "DQ Alpha" or DQ ] is a sequence variant
locus. This locus codes for production of a protein. NRC 2 at 71. DQ
Alpha has eight alleles, although the tests used only detect six
alleles. Ibid. at 74. With six alleles, there are 21 possible genotypes:
six homozygous and 15 heterozygous. Ibid. at 71. At the time of the
first NRC report in 1992, only the DQ Alpha locus was used for forensic
testing. NRC 1 at 70. Test kits were purchased from the Cetus
Corporation. These kits contained the primers, polymerase and typing
strips necessary to conduct DQ Alpha typing of a sample. The San
Francisco Police Crime Lab ("Crime Lab") does not use this
kit, but uses a newer kit manufactured by Perkin-Elmer that
simultaneously types the DQ Alpha locus with five other loci, known
together as the Polymarkers.
Instead of amplifying only one gene, like the DQ Alpha test, the
polymarker ("PM") test amplifies five loci simultaneously:
LDLR (low-density lipoprotein receptor), GYPA (glycophorin A), HBGG
(hemoglobin gamma globin), D7S8 (a marker on chromosome 7) and GC
(group-specific component). Each of the loci are bi- or tri- allelic,
totaling 972 genotypic combinations. NRC 2 at 74. The first forensic
polymarker tests were conducted with kits designed only to amplify the
polymarker loci. As mentioned above, the Crime Lab uses a newer kit
which amplifies and types not only the polymarker loci, but also the DQ
Alpha locus.
In addition to typing DQ Alpha and polymarker loci, the Crime Lab also
types the D1S80 locus. D1S80 is a length-variant locus classified as a
VNTR. The base sequence of D1S80 is 16 bases long. There are more than
30 alleles. Because D1S80 varies by length, the alleles are differences
in the number of times the base sequence repeats. NRC 2 at 72. D1S80
does not confirm the sequence of the repeating bases, and "some
ambiguous alleles are caused by insertion or deletion of a single base
…." Id.
The D1S80 locus is typed using electrophoretic separation in a technique
called PCR-amplified fragment length polymorphism technique (AMP-FLP.).
This technique has characteristics of both PCR and RFLP typing. After
the fragments are amplified using PCR, their lengths are determined
using a technique called electrophoresis. Electrophoresis is a process
where the amplified sample is placed on one end of a gel. An electric
current is applied to the gel. The DNA fragments move across the gel;
the shorter fragments move faster than the longer fragments. At the end
of the process the fragments, which appear as bands, are lined up along
the gel according to size. The length of the fragments are compared to a
control sample, known as an allelic ladder. The determination of the
actual length of the unknown samples is a subjective determination made
by the technician conducting the analysis.The San Francisco Crime Lab
began PCR-based DQ Alpha + Polymarker and D1S80 typing in February,
1997, but did not issue its first formal report using these tests until
April, 1997. The DQ Alpha+Polymarker kit was used in this case in
June,1998. Although the D1S80 kit has not yet been used in this case, it
is anticipated that such results are forthcoming.
Yet another type of length-variant loci used for forensic typing by the
Crime Lab are Short-Tandem Repeat ("STR") loci. These loci are
2 to 7 base pairs long. Bernd Brinkmann, Overview of PCR-Based Systems
in Identity Testing in Forensic DNA Profiling Protocols (Patrick J.
Lincoln and Jim Thomson, eds. 1998) 106. Like D1S80, the STR loci vary
by length, so typing the loci do not reveal the base sequence of the
locus. However, unlike D1S80,and unlike some early versions of STR
testing, in which one locus is typed by means of gel electrophoresis and
visualized by staining, the STR loci were typed in this case by a true
"black box" technology in which several loci are run
simultaneously through a automated capillary electrophoresis system and
then typed not by a human, but by complicated computer analysis which
uses a flourescent-based laser detection system. In March, 1998, The
Crime Lab began typing three STR loci ( THO1, TPOX, and CSF), and a sex
determination of the amelogenin gene, using a "Green I" typing
kit, a 310 Genetic Analyzer, and a fully automated capillary
electrophoresis system. It was this typing system that Superior Court
Judge Robert Dondero recently ruled in People v. Bokin et al., SC.
No.168461 did not meet Kelly- Frye standards. A copy of Judge Dondero's
written seventeen page opinion, issued after an extensive five month
hearing, is attached to this pleading as exhibit A.
Finally, as part of their flourescent-based STR testing, the Crime Lab
in February 1999 began to use a new STR typing kit called "Profiler
Plus", which types nine additional STR loci
( D3S1358, vWa, FGA, D8S1179, D21S11, D18S51, D5S818, D13S317, D7S820).
The present case is apparently the first time the Crime Lab has used
this kit in actual casework. Significantly, in December, 1998 a
federally funded Audit Team conducted a long delayed audit of the Crime
Lab, and on April 7, 1999 the Audit team issued a highly critical report
detailing deficiencies in many areas. Importantly, the report concludes
that " (t)he deficiencies observed should be addressed before the
program purses casework DNA typing involving the Profiler-Plus and
Cofiler systems." A copy of the Audit Report, as well as a letter
from the Crime Lab responding to the Report, is attached hereto as
Exhibit B.
Once an evidentiary sample collected from a crime scene, and blood from
the victim and suspect(s) has been extracted and the genotypes of each
determined, the technician will calculate a frequency for the occurrence
of a given genotype in either the general population or within a given
racial or ethnic population. In other words, the frequency of a given
allele in a random collection of samples is used to calculate the
probability of the occurrence of a particular genotype or collection of
alleles. One expert describes this number as the "random match
probability … a theoretical likelihood that a randomly selected person
from the general population (of from the population of certain large
ethnic or racial groups) would genetically match the trace evidence as
well as the defendant." In this case, the Crime Lab claims in its
report that the random match probabilities are as follows: 1 out of 19
thousand trillons for Caucasians, 1 out of 280 thousand trillions for
Blacks, and 1 out of 340 thousand trillions for Mexican- Americans.
Based on the genotypes determined, the technician will use a database
for the particular marker or groups of markers, to determine the
frequency of the genotype in the population. The databases contain
frequencies of genotypes. To illustrate, if the DQ test of the suspect's
blood gives a type of 1.1, 2, the technician will chose the frequencies
of 0.036 for Caucasians, 0.036 for African Americans, and 0.022 for
Hispanics. The inverse of each number is taken to determine the
occurrence of the genotype in the population, in this example, 1 in 28
for Caucasians and African Americans, and 1 in 45 for Hispanics.
When more than one marker is tested such as in the PM test or when the
PM test is combined with the DQ test, the initial frequency calculation
is done for the alleles at each gene or marker. The frequencies for each
gene are then multiplied together using the Product Rule, and an overall
frequency is determined. The inverse of this number is then taken and
the overall frequency is found. Again, to illustrate, a suspect has the
genotype of DQ 1.1, 2 and the LDLR of A, A. The calculations are as
follows:
Caucasian African American Hispanic
DQ 0.036 0.036 0.022
LDLR 0.23 0.08 0.23
LDLR x DQ 0.008 0.003 0.005
Inverse 1 in 125 1 in 333 1 in 200
As shown above, the addition of each marker will increasingly narrow the
number of individuals within the population that will have the given
genotype. However, these tests do not and cannot show whether the
suspects blood is the blood present in the unknown sample, even if the
suspect's genotype matches that of the unknown sample. The only
conclusive result from any of the tests conducted in this case is to
exclude an individual.
The validity of using the Product Rule rests on a number of assumptions.
First, the population studied engages in random mating--that mates are
not chosen based on their genotype at the markers used in DNA testing.
NRC 2 at 90. If the alleles of a gene are subject to random mating, then
the alleles should be in what is called Hardy-Weinberg (HW) equilibrium.
A population is in HW equilibrium if the actual observations of each
genotype are consistent with that expected under the HW test. Id. at
91-92. A population will not be in equilibrium if the occurrence of a
particular genotype in a population is significantly different than
expected after statistical analysis. The lack of equilibrium indicates
that there is some substructure to the population so that the alleles of
a given gene are not assorting independently from generation to
generation.
When multiple genes are used, as they are in this case, an additional
assumption must be made, that is that the loci are in linkage
equilibrium. NRC 2 at 106. Linkage equilibrium is the point at which the
frequency of a multi-locus genotype is equivalent to the product of the
genotype at each loci in a randomly mating population. Id. In other
words, the frequency of a particular genotype involving more than one
gene or marker, is equal to the frequency of one marker multiplied by
the frequency of the second marker.
To summarize, to conduct bloodtyping analysis using one of the above
methods, the following steps must be followed. First, a sample
containing human DNA must be extracted from a forensic sample. The DNA
must be amplified using PCR. The amplified sample is then typed for
alleles at specific loci. Once the sample has been typed, the results
are compared to a known sample of DNA, which has also been typed at the
same loci, using the same techniques. If there is a reported match,
then, using a database of genotype frequencies, the analyst determines
the random match probability that the known sample and unknown sample
are from the same source.
D. THE SCOPE OF KELLY-FRYE
The prosecution seeks to introduce DNA evidence derived from pooling of
fingernail scrapings which were collected over 12 years ago, which
allegedly contain DNA profiles of several people, and which was derived
from four commercially developed test kits: the DQ Alpha + Polymarker
kit, the D1S80 kit, the Green I kit, and the Profiler-Plus kit. The
particular kits and methodologies used in this case are novel: they have
never been reviewed for admissibility by an appellate court in
California . Even the best established methods (those based on RFLP
analysis) are controversial and have recently been held inadmissible by
the Supreme Court of California. People v. Venegas (1998) 18 Cal.4th
47(reversal of rape conviction is required for lack of compliance by the
FBI with procedures recommended in 1992 by the National Research Council
(NRC) for determining the statistical probability of a random match).
The newer PCR-based systems have produced conflicting opinions in the
trial courts. In the Riverside county case of People v. Scott, S. C. No.
CR-48638, Judge Bailey conducted an extensive evidentiary hearing before
concluding in February, 1997 that the D1S80 typing kit did not meet
prong one of the three part Kelly-Frye standard. A copy of his ruling is
attached as exhibit C. In People v. Bokin, Judge Dondero found that the
faults with the D1S80 kit identified in Scott could be addressed under
prong three of the three prong test of Kelly, but he also ruled that the
Green I STR typing kit did not meet Kelly -Frye standards under prong
one because it was not shown that it was generally accepted in the
broader scientific community, which includes not only forensic
scientists, but also molecular biologists. In a third case, People v.
McClanahan, S. C. No. 162412, Judge Kramer considered the Green I kit
but not D1S80, and he ruled that the Green I kit met prong one because
it was generally accepted by forensic scientists. Needless to say, the
evidentiary use of PCR- based testing systems is controversial.Cf.,
People v. Soto (1999) 21 Cal. 4th 512, 535-539 (reviewing the ebb and
flow of scientific and judicial opinion regarding RFLP testing).
According to critics of DNA testing procedures, "the major lesson
of the controversy is that forensic laboratories are lacking in
scientific rigor and that new forensic techniques receive inadequate
scientific scrutiny before they are presented in court. By this account,
an atmosphere of hostility surrounds the issue because promoters of
forensic DNA tests have responded to legitimate criticisms by attacking
the critics rather than heeding their message."Thompson, Evaluating
the Admissibility of New Genetic Identification Tests: Lessons From the
'DNA War', 84 J. of Crim. Law & Criminology 22
Nowhere is this hostility better illustrated than in the recent
intemperate and unfair remarks of DNA expert Ed Blake, who in 1996 was
awarded a lucrative city contract to supervise the San Francisco Crime
Lab's DNA Program. Lashing out at the recent ruling of Superior Court
Judge Robert Dondero in People v. Bokin, which was highly critical of
the SFPD's DNA Program, Blake has called for the removal of Judge
Dondero from office on the outrageous ground that the judge is "an
arrogant, ignorant, stupid judge who is unable to get through the smoke
thrown at him by a very skilled lawyer.""Judge Rejects DNA
Test Evidence", San Francisco Examiner, May 7, 1999, p1; see also,
Comments of Professor William Thompson, People v. Bokin: An Internet
Discussion, http://CA_v_Bokin.pdf@www.scientific.org ("We will get
a lot farther through respectful exchanges...than we will by publically
declaring that those who disagree with us are ignorant, stupid or
callous.") Ironically, Blake is the so-called "impartial"
expert that gave crucial, but misleading (see infra at pp. 38-40)
testimony in People v. Morganti (1996) 43 Cal. App. 4th 643, a case the
People ask this Court to follow here.
Prosecutors and commercial vendors of DNA test kits typically portray
forensic DNA testing as a science that admits of no error in the
identification of individuals. See e.g., "Judge Rejects DNA Test
Evidence", San Francisco Examiner, May 7, 1999, p1 ( " Frank
Gannon, Foster City public relations manager for the manufacturer, said
his firm was confident the ruling [of Judge Dondero] would be
overturned. ‘Our tests show this test 100 percent accurate,'he said.
‘We have never had a mistake. We've never put someone at the scene of
a crime who was not there. It's impossible.'") ; Rockne P. Harmon,
"‘Barney' the Dinosaur: While Court Dawdles, DNA Admissibility
Standard Fossilizes, San Francisco Daily Journal, June 8, 1998, p.4
(Alameda County prosecutor claims that all DNA testing methods are
infallible and that the decision in Venegas and the Court's inaction in
other pending DNA cases means that "clarification of four years of
positive Court of Appeal DNA opinions is unlikely until long after the
November election in which three of the Supreme Court justices are up
for confirmation.")
Prosecutors and manufacturers views, however, do not generally reflect
an understanding of the several sciences that comprise the foundations
of forensic DNA analysis. The techniques that are employed in forensic
DNA laboratories represent a blend of the old and the new in molecular
biology, chemistry, population genetics, and statistical science. The
fact that various elements of the new technology have been used
successfuly in other applications for some years does not mean that
those elements can be combined in a manner that will result in a fully
matured science at the outset. Yet that is how DNA testing has been
portrayed since its initial appearence at criminal trials in the late
1980's. Fortunately, our Supreme Court, unlike some others, has not been
so easily misled.
1. The Admissibility of DNA Evidence Must Be Determined Under The
Standard of People v. Kelly.
The admissibility of each item of DNA evidence depends on whether it
was derived from a method that is generally accepted to be reliable. To
make this determination, the court must apply the standard set forth in
People v. Kelly (1976) 17 Cal.3d 24, 130 Cal.Rptr 144. The Kelly
standard has three "prongs":
(1) it must be established, usually by expert testimony, that the
scientific methods utilized are generally accepted as reliable by the
relevant scientific community,
(2) the witness furnishing such testimony must be properly qualified as
an expert to give an opinion on the subject, and
(3) the proponent of the evidence must demonstrate that correct
scientific procedures were used in the particular case.
Kelly, 17 Cal.3d at p. 30 [emphasis in original].
Importantly for the present case, Venegas recently clarified that
"the Kelly test's third prong does not apply the Frye requirement
of general acceptance -it assumes the methodology and technique in
question has already met that requirement. Instead, it inquires into the
matter of whether the procedures actually utilized in the case were in
compliance with that methodology and technique, as generally accepted by
the scientific community. ...The third prong inquiry is thus case
specific; ‘it cannot be satisfied by relying on a published appellate
decesion.'" 18 Cal 4th at 78 (emphasis added).
In People v. Farmer (1989) 47 Cal. 3d 888, 913 the Court had stated that
" careless testing affects the weight of the evidence and not its
admissibility..." However, in Venegas the Court clearly retreated:
" Our reference to ‘careless testing affecting the weight of the
evidence and not its admissibility' in Farmer ...was intended to
characterize short-comings other than the failure to use correct,
scientifically accepted procedures such as would preclude admissibility
under the third prong of the Kelly test." 18 Cal. 4 th at p.80
(emphasis in original). See also, People v. Soto (1999) 21 cal. 4th 512,
519 ("The proponent of the evidence must also demonstrate that
correct scientific procedures were used.")
As Venegas teaches, the goal of the Kelly standard as applied to DNA
evidence is to "forestall the jury's uncritical acceptance of
scientific evidence or technology that is so foreign to everyday
experience as to be unusually difficult for laypersons to evalute."
18 Cal. 4th at p.80. "DNA evidence is different", and "
is an instance in which the method of scientific proof is so
impenetrable that it would ‘...assume a posture of mystic
infallibility in the eyes of a jury...'"18 Cal.4th 47, 81-84. There
is also concern that efforts to challenge controversial scientific
evidence through cross-examination and expert testimony will consume
inordinant amounts of time and raise issues "beyond the scope of
critical analysis by the average lay person." In re Amber B. (1987)
191 Cal.App.3d 682, 690-91.; Accord People v. McDonald 37 Cal.3d 351,
372 (1984). The greatest danger arises when a technique carrying an
"aura of infallibility" is, in fact, quite fallible, and when
an analysis of the reliability of the technique involves technical
issues that are difficult for lay individuals to understand. Again, DNA
evidence " is an instance in which the method of scientific proof
is so impenetrable that it would ... assume a posture of mystic
infallibility in the eyes of a jury." 18 Cal. 4th at 83-84
The role of the Court, when applying the general acceptance test to DNA
evidence, is to determine "the existence, degree, (and) nature of a
scientific consensus or dispute (with) the interpretative assistance of
qualified live witnesses subject to a focused examination in the
courtroom."People v. Soto (1999) 21 Cal. 4th 512, 540-541 n. 31 The
court must " conduct a 'fair overview' of the subject, sufficient
to disclose whether 'scientists significant either in number or
expertise publicly oppose [a technique] as unreliable.People v. Reilly
(1987) Cal.App.3d 1127, 1148, quoting from People v. Brown (1985) 40
Cal.3d 512, 533.
As Venegas teaches, "(i)n determining the question of general
acceptance, courts must consider the quality, as well as quantity, of
the evidence supporting or opposing a new scientific technique." 18
Cal. 4th at 85(emphasis added) And as Venegas also makes clear, the
court's role in the inquiry is not one of abdication to the scientists.
Thus, in Venegas, the Court adopted the modified ceiling approach
advocated in the 1992 NRC report, notwithstanding that the 1996 NRC
report, as well as numerous other scientific articles authored by
well-credentialed experts, and testimony from such experts at the Kelly
hearing, all assailed the approach as scientifically unsound. 18 Cal 4th
at 86-89. More recently, in People v. Soto, supra, the Court emphasized
that "in a context of rapidly changing technology, every effort
should be made to base (decision) on the very latest scientific
opinions..." 21 Cal. 4th at 540-541 n. 31 The point here is that
the court's role in the inquiry is not one of a potted-plant.
2. Kelly Requires That Each Major Step of a DNA Typing Procedure Be
Generally Accepted as Reliable by The Scientific Community
In reviewing the admissibility of DNA evidence under Kelly, this
court does not write on a blank slate. Several California appellate
rulings on DNA evidence have helped establish the nature and scope of
the issues reviewable under Kelly. See, People v. Soto, supra; People v.
Venegas, supra; People v. Allen (1999) 72 Cal. App. 4th 1093; People v.
Wright (1998) 62 Cal. App.4th 31; People v. Morganti (1996) 43 Cal. App.
4th 643; People v. Axell (1991) 235 Cal.App.3d 836 ; People v. Barney
(1992)8 Cal.App.4th 798; People v. Wallace (1993)14 Cal.App.4th 651.
Because of the importance in this case of Allen,Wright and Morganti,
these decisions are discussed in section IIB infra. One key point
established by the other cases is that forensic DNA testing requires a
series of distinct steps, and that the methods employed at each major
step are independently reviewable under Kelly; each step must be
generally accepted. For example, Soto, Venegas, Axell and Barney
identified three major steps in DNA tests that employ RFLP analysis and
made it clear that failure to use a generally accepted method for any of
these steps would preclude admissibility of the resulting evidence. For
newer technologies, this court will need to identify the steps in the
testing procedure and ascertain that a generally accepted method is
being used at each step, particularly at critical steps (i.e., points
where serious errors might occur).
Statistical computation will obviously be a critical step for every DNA
testing procedure that purports to find a "match" between
samples with respect to their genetic characteristics. See People v.
Venegas 18 Cal 4th at 82 ( " A determination that the DNA profile
of an evidentiary sample matches the profile of a suspect establishes
that the two profiles are consistent, but the determination would be of
little significance if the evidentiary profile also matched that of many
or most other human beings."); People v. Barney, 8 Cal. App. 4th at
803 ("[t]he statistical calculation step is the pivotal element of
DNA analysis, for the evidence means nothing without a determination of
the statistical significance of a match of DNA patterns."); Axell,
supra, 235 Cal.App.3d 836, 866 ("We find that...a match between two
DNA samples means little without data on probability..."); People
v. Wallace (1993) 14 Cal. App. 4th 651 n. 3 (1993) (without valid
statistics DNA evidence is "meaningless").
Recent scientific commentary has clarified the necessary elements of the
statistical estimation procedure. In particular, it is now clear that no
statistical procedure is acceptable unless it addresses the probability
of two events that could cause a "match" to be reported
between samples from different people: (1) a coincidental match between
different individuals who happen to have the same genetic
characteristics, and (2) a false positive (false match) due to
laboratory error. The probability of a coincidental match is typically
estimated by determining the frequency of matching genetic
characteristics (genotypes) in a suitable reference population (or
populations). The probability of a false positive is estimated by
determining the laboratory's rate of errors in proficiency tests. See
generally, L. Mueller, The DNA Controversy And NRC II, in Statistical
Methods in the Health Sciences: Genetics, ME Halloran & S. Geisser
(eds., 1999); D.H. Kaye, DNA, NAS, NRC,DAB, RFLP, PCR, and More: An
Introduction to the Symposium on the 1996 NRC Report on Forensic DNA
Evidence, (1997) 37 Jurimetrics: The Journal of Law, Science, and
Technology 395.
Another critical step for every DNA testing procedure is the exact
methodology used to determine the various DNA types within each testing
system. In People v. Venegas, (1998) 18 Cal. 4th 47, the Supreme Court's
discussion of prong one did not focus on a particular locus or even on a
generic methodology such as RFLP, but rather on the "general
scientific acceptance of the FBI's RFLP methodology used in this
case...." See also, People v. Brown (1985) 40 Cal. 3d 512, 528, 534
("[T]he prosecution failed to satisfy the Kelly/Frye rule by
demonstrating at trial the scientific acceptance of the tests
performed." Moreover, the issue could not be resolved on appeal
because " the electrophoretic method itself is apparently performed
under substantial chemical and electrical variations, and considerable
training and experience are necessary to interpret the visual
results."); People v. Wright , (1998) 62 Cal. App. 3d 31, 38 ( PCR
methods used to analyze DQ alpha and polymarker loci were admissible in
part because witnesses testified that " the PCR technique, and the
DQ Alpha and polymarker methods used to conduct the PCR analysis in this
case, are both generally accepted as reliable and valid in the
scientific community."); People v. Barney (1992) 8 Cal. App. 4th
798, 824 ( " Moreover, the survival of the third-prong inquiry is a
matter of common sense. If it is not established that correct scientific
procedures were used in the particular case, it cannot be known whether
the test actually conducted was the one that has achieved general
scientific acceptance."); People v. Axell (1991) 235 Cal. App. 3d
836, 849 (" After reviewing the protocols used by Cellmark labs and
laboratory notes in this case, Doctor Roberts [and six other scientists
were] 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."); People v. Taylor (1995)
33 Cal App. 4th 262, 266 ("Kelly's test of general acceptance in
the scientific community applies not only to the techniques for
laboratory analysis of the DNA material but also to the statistical
calculations which indicate the likelihood that the crime scene sample
of DNA would match with some individual other than the defendant.... At
the time of the Kelly hearing conducted in this case in February l993
there was authority for the proposition that the laboratory processes
and protocols used in DNA analysis by Cellmark Diagnostic-the testing
company which performed the test conducted in this case-met with general
scientific acceptance.")
The focus on the specific methods and procedures used in the particular
case is universal in Kelly- Frye litigation. As stated in State v.
Jackson (Neb. 1998) 255 Neb. 68, 582 N. W. 2d 317, 325 a case in which
the Nebraska Supreme Court recently ruled that the results of an
unspecified STR procedure should not have been admitted absent a
foundation that the lab had followed its own testing protocols, the
admissibility of PCR- based testing methods is generally held to turn on
the following issues:" (1) whether the witnesses on the DNA issue
are experts in the relevant scientific fields, (2) whether the type of
DNA testing used in the case under consideration is generally accepted
by the relevant scientific communities, including the accompanying
statistical analysis; (3) whether the testing protocol used in the case
under consideration is generally accepted as reliable if performed
properly, (4) whether the test conducted properly followed the protocol;
(5) whether DNA analysis evidence is more probative than prejudicial ...
and (6) whether statistical probability evidence interpreting the DNA
analysis results is more probative than prejudicial."See also,
United States v. Beasley (1996 8th Cir.) 102 F. 3d 1440, 1445 ("[T]he
PCR method of DNA typing using the DQ alpha Amplitype test kit and the
Polymarker test kit has achieved general acceptance within the forensic
science community."); United States v. Lowe (D. Mass 1997) 954 F.
Supp. 401, 411(Changes in the FBI's RFLP methodology are significant
enough to require Daubert analysis even though general RFLP methodology
has already been ruled admissible);United States v. Gaines (S.D. Fla
1997) 979 F. Supp. 1429,1437 (" For DQA1 and PM, the process used
for determining a match involves the use of the reverse dot blot test.
For D1S80 and Amelogenin, the process used is gel electrophoresis. Since
the processes are different, they will be addressed
separately.")Commonwealth v. Blasioli, (1998) 552 Pa. 149, 1998 WL
313388 ("This Court has generally required that both the theory and
technique underlying novel scientific evidence must be generally
accepted.") Commonwealth v. Sok, (Mass. 1997) 425 Mass. 787,683
N.E. 2d 671,798 ("Here we are primarily concerned with whether the
particular test kits were reliable when utilized by CBR."); State
v. Harvey, (N.J. 1997) 151 N.J. 117, 699 A. 2d 596,622-623
( " Proving general acceptance ‘ entails the strict application
of the scientific method, which requires an extraordinarily high level
of proof based on prolonged , controlled, consistent, and validated
experience.'...Here, the State's burden is to prove that the polymarker
test and the interpretation of its results are non-experimental,
demonstratable techniques that the relevant scientific community widely,
but perhaps not unanimously, accepts as reliable.")( acceptance of
PM generally did not preclude challenge to dot -intensity analysis, as
" a court must examine each step of a scientific process or
technique."); People v. Dalcollo (1996) 282 Ill. App.3d 944,669
N.E.2d 378( " [W]e hold that the Frye test requires that both the
theory and the techniques or procedures implementing the theory must be
generally accepted in the relevant scientific
community.")(collecting numerous out of state cases in support)
Franson v. Micelli (1994) 269 Ill. App. 3d 20, 645 N.E. 2d 404, vacated
on other grounds, 666 N.W. 2d1188 ( "The first reason we reject the
Lipscomb/Miles cases is because these courts improperly applied the Frye
test. The Lipscomb court reasoned that Frye applies to new scientific
principles and the principle involved is the general DNA forensic
analysis and not the procedures used within that framework. Thus, the
court asserted that the fact that different companies may use different
procedures in arriving at their results only goes to the weight of the
evidence. The first flaw in this reasoning is that, although ‘general
DNA forensic analysis' may be the new scientific principle, Frye
requires that ‘the thing from which the deduction is made' must be
generally accepted. It cannot be disputed that the theory behind DNA
testing is well accepted. This does not mean, however, that simply
because the concept is recognized the procedures used are not subject to
a Frye analysis and any challenge to their validity only goes to the
weight of the evidence. Under Frye, if the procedure or ‘thing' upon
which the DNA result was determined is not generally accepted, then the
result is inadmissible.") Along with these courts and the first NRC
Report, the defense urges the Court to hold that " (i)t is
meaningless to speak of the reliability of DNA typing in general-without
specifying a particular method." NRC I at p. 8.(emphasis added)
In the recently conducted Kelly-Frye hearings before Judge Dondero in
People v. Bokin, the People's formulation of the prong one issue was
improperly focused on " STR's and D1S80", which are not
methods at all, but rather a "class of repeated
units"(STR's)(NRC II, p. 70) and a locus(D1S80) (NRC II, p. 72).
The method used in this case to detect alleles at the DQ Alpha and
polymarker is a reverse-dot blot method utilizing the DQ Alpha +Polymarker
test kit. The method used in this case to detect alleles at the D1S80
locus is gel electrophoresis utilizing the Perkin Elmer D1S80 kit. The
method used in this case to detect alleles at the thirteen STR loci is
automated capillary electrophoresis utilizing the 310 Genetic Analyzer
and the Green I and Profiler-Plus kits. The method used to make
frequency calculations in this case is the unmodified product rule
derived from the particular databases used by the San Francisco Police
Department. These testing methods, not PCR generally (see, United States
v. Hicks (9th Cir. 1996) 103 F. 3d 837,844 ["The PCR method itself
is not a genetic test; it is a mere amplification technique."], and
certainly not the loci themselves, should be the proper focus of prong
one.
The failure to grasp this distinction led the People astray from the
outset in Bokin and in People v. McClanahan, before Judge Kramer. For
example, in Bokin, instead of having their sole expert Dr. Holt review
the protocols used and render an opinion about the general acceptance of
the specific procedures outlined in the protocols, as was done in Axell
and many other cases, the People contented themselves with leaving the
witness in the dark about the specific procedures used by the San
Francisco Crime Lab (see, R.T. 394) and asking such vague and
meaningless questions as " Are the STR loci THO1, TPOX,...and
CSF1PO...generally accepted in the scientific community as a reliable
means of human identification?" (R.T. 45) At this level of
generality, and without specific definition of either the relevant
scientific community or the specific methods being utilized, the answer
is predictable, but irrelevant, much as it would be in a hypothetical
case in which a group of psychics, thinking they are scientists and
believing they can make valid fingerprint comparisons by holding the
known and questioned prints to their turbaned heads, are asked the
question, "Is fingerprinting generally accepted in the scientific
community as a reliable means of human identification?" The
People's presentation in both Bokin and McClanahan was that meaningless.
3. The Court Must Look Beyond Forensic Science For Evidence of General
Acceptance.
The general acceptance test of Kelly cannot be met by showing that
promoters and practitioners of the method accept it to be reliable. The
test is not whether a method is accepted by those who have a personal or
professional stake in its acceptance, but rather, whether it is
"accepted as reliable by the larger scientific community in which
it originated." People v. John W. 185 Cal.App. 3d 801, 805 (1986);
People v. Shirley 31 Cal.3d 18, 54 (1982).
Courts have also recognized that promoters and practitioners of a
particular method "may be too closely identified with the
endorsement of [the technique] to assess fairly and impartially the
nature and extent of any opposing scientific views." Kelly, supra,
17 Cal.3d at 38. Thus, when applying the Kelly standard, the court must
look to experts who are "'impartial,' that is, not so personally
invested in establishing the technique's acceptance that he might not be
objective about disagreements within the relevant scientific
community." People v. Brown, supra, 40 Cal.3d 512, 530; accord
People v. Venegas, 18 Cal. 4th at 77 (" FBI Agent Lynch, though
vigorously defending the merits of the FBI's RFLP analytical procedures
she had followed in this case, did not purport to be qualified, as a
molecular biologist or otherwise, to testify on questions of general
scientific acceptance of the validity of those prosedures.") In
this regard, the court should bear in mind that employees of forensic
labs "have a clear pecuniary interest in the acceptance of DNA
evidence by the courts. The success of their employers and the stability
of their own employment depends upon continued use of DNA testing."
Dan L. Burk, DNA Identification: Possibilities and Pitfalls Revisited,
31 Jurimetrics J. 53, 79-80. Obviously, employees of laboratories that
hold patent rights to a new forensic technique have a similar pecuniary
interest in the acceptance of the technique in court.
As Judge Dondero recently held in Bokin, the relevant scientific
community for purposes of the prong one inquiry is the broader community
of molecular biologists and population geneticists and not simply the
community of forensic scientists using DNA technology, as was held by
Judge Kramer in McClanahan. This principle is dictated by relevant case
law. See, People v. Soto, 21 Ca. 4th at 515 (Issue is whether the RFLP
statistical methodology " is generally accepted in the relevant
scientific community of population geneticists."); People v.
Venegas 18 Cal. 4th at 77 ("FBI Agent Lynch...did not purport to be
qualified, as a molecular biologist or otherwise, to testify on
questions of general scientific acceptance of the validity
of...procedures.");People v. Axell (1991) 235 Cal. App. 3d 836,
857("The search for scientific consensus must be from within ‘the
particular field in which it belongs'. Since DNA profiling is an
amalgamation of primarily two disciplines, molecular biology and
population genetics, it appears logical to consider its acceptance by
those communities for forensic use."); People v. Reilly (1987) 196
Cal. App. 3d 1127, 1138-1139(same).
The failure of the People to recognize this principle during the course
of two hearings resulted in inadequate evidentiary showings which
myopically focused solely on the views of the forensic community. In
McClanahan, Dr. Holt's entire testimony was strictly limited to whether
she felt that "STRS" were generally accepted by "the
forensic DNA community."(R.T. 118; 127,133,159,260). D1S80 was not
an issue in that case. In Bokin, the People again relied exclusively on
Dr. Holt and had her testify that "STR and D1S80 typing" are
generally accepted in the "relevant scientific community".(R.T.
44) Although that community was never explicitly defined, it is clear
from the overall context of her answers and her emphasis on the annual
gathering of forensic DNA analysts at the Promega Proceedings, and on
the widespread use of STR methods by law enforcement personnel, that she
was once again talking about the forensic science community. Indeed, at
one point in cross examination she acknowledged this narrow focus (R.
T.405), she admitted that " I read mostly forensic journals",
and she conceded she was not current on the literature of the criticisms
within the medical community, outside of forensics, about the use of
STR's and PCR. (R.T. 295-296). The same narrow focus was evident in
McClanahan, where Judge Kramer evidently and erroneously accepted the
view that the relevant scientific community could be restricted to the
forensic community. By contrast, Judge Dondero's mid-hearing ruling that
the relevant scientific community extended beyond forensic scientists
gave the People fair warning that they had to broaden the scope of their
prong one presentation. They never did, with the result that the Bokin
record contained a glaring failure of proof on prong one issues, a
failure made most conspicuous by a comparison with other cases in which
prosecutors have successfully established the general acceptability of
other DNA testing methods. See e.g., People v. Soto, 21 Cal. 4th at 527
(" During the 11 days of Kelly hearing testimony... the parties
elicited extensive DNA testimony from 3 prosecution experts: the
criminalist, Robert Keister, and two human population geneticists,
Ranajit Chakraborty and Bruce Kovacs."); People v. Axell (1991) 235
Cal. App. 3d 836, 849,856 (" At the hearing, the People presented
four expert witnesses in their case- in- chief and two in
rebuttal.")
The history of litigation on forensic DNA tests shows the wisdom of
looking beyond forensic science to the "larger scientific
community" to determine what is and is not accepted. When RFLP-based
DNA evidence was first introduced in U.S. courts in 1987, it faced
almost no opposition. Few scientists, other than those who were
employees or consultants of the forensic laboratories knew the details
of forensic testing procedures, and few had occasion to think about the
difficulties posed by the transfer of DNA technology to the forensic
arena.
Forensic DNA tests were ... developed by commercial laboratories whose
procedures were proprietary, whose work product is generally not
available for examination by outsiders (except when used in a court
hearing), and whose laboratory protocols were initially available only
by court order. The forensic scientists who developed and validated the
techniques in both commercial and government laboratories were well
aware that they would be called upon to defend the new tests when their
admissibility was challenged in litigation, and they had a professional
stake in the outcome of that litigation. In that atmosphere, they had
little incentive to air and discuss problems, and strong incentive to
suppress doubt and uncertainty for fear that candid statements would be
used against them in court.
Thompson, Evaluating the Admissibility of New Genetic Identification
Tests: Lessons From the 'DNA War', 84 J. of Crim. Law & Criminology
22, at 95.
During the first few years that these tests were used, there was little
hint of the "raging controversy" that would later arise,
leading to their exclusion in Venegas, Barney and Wallace. Indeed,
nearly two years passed before the first scholarly critique of forensic
DNA test procedures appeared in the scientific literature. Eric Lander,
DNA Fingerprinting on Trial, 339 Nature 501 (1989).
The absence of scholarly criticism did not mean that the methods used by
forensic DNA laboratories were "generally accepted." Many
laboratories were, in fact, using methods that have since been rejected
by the scientific community.
Without the benefit of open scientific scrutiny, some testing
laboratories initially used methods (for such fundamental steps as
identifying patterns, declaring matches, making comparison with a
databank, and correcting for band shifting) that they later agreed were
not experimentally supported.
NRC 1 at 56.
Before the methods could become controversial, they had to become known
and understood. The absence of criticism when the tests were first
introduced was not proof that the tests would be generally accepted in
the scientific community; silence does not imply assent. Thus, to
conclude that a technique is "generally accepted" when only a
few scientists (other than its promoters) have evaluated it, is perilous
indeed.
4. Neither Acceptance of PCR For Nonforensic Scientific Purposes Nor
Widespread Law Enforcement Use of PCR Testing Kits For Databasing and
Other Nonevidentiary Purposes Establishes That Forensic Tests Employing
PCR Are Reliable or Generally Accepted For Courtroom Use
The People in McClanahan and Bokin made the argument that forensic
DNA tests must necessarily be accepted in the scientific community
because they employ procedures (e.g., PCR) that are used and accepted
elsewhere in science for other purposes. Their argument is syllogistic,
viz.: PCR analysis is accepted; SFPD's Crime Lab uses PCR; therefore the
Lab's testing in all its variations is accepted. The problem with this
argument is that it fails to recognize the difficulties that may arise
from the transfer of technology from one application to another. It is
widely recognized that forensic DNA testing is more technically
demanding than other applications of RFLP and PCR technology, and that
it involves additional critical steps that do not arise in other
applications (such as the matching and statistical estimation steps).
What must be considered under Kelly is whether the method that produced
the incriminating evidence is accepted to be reliable as it was applied,
not whether a similar method is accepted for another purpose.
The simpleminded notion that any forensic DNA test employing RFLP or PCR
is, of necessity, reliable and accepted was flatly rejected by the
National Research Council:
Before a method can be accepted as valid for forensic use, it must be
rigorously characterized in both research and forensic settings to
determine the circumstances under which it will and will not yield
reliable results. It is meaningless to speak of the reliability of DNA
typing in general--i.e., without specifying a particular method.
NRC 1 at 51-52.
Moreover, even as to forensic use of a particular method, our Supreme
Court has explicitly rejected "widespread use" by law
enforcement as a surrogate for a searching inquiry into whether
impartial scientists accept a particular technique. In People v. Leahy
(1994) 8 Cal. 4th 587, 605-606, the Court stated unambiguously :
The People observe that HGN testing has been used by law enforcement
agencies for more than 30 years....
In determining whether a scientific technique is "new" for
Kelly purposes, long-standing use by police officers seems less
significant a factor than repeated use, study, testing and confirmation
by scientists or trained technicians...To hold that a scientific
technique could become immune from Kelly scrutiny merely by reason of
long- standing and persistent use by law enforcement outside the
laboratory or the courtroom, seems unjustified.
The United States Supreme Court agrees: "Respondent argues that
because the Government--and in particular the Department of
Defense--routinely uses polygraph testing, the Government must consider
polygraphs reliable. Governmental use of polygraph tests, however, is
primarily in the field of personnel screening, and to a lesser extent as
a tool in criminal and intelligence investigations, but not as evidence
at trials.... Such limited, out of court uses of polygraph techniques
obviously differ in character from, and carry less severe consequences
than, the use of polygraphs as evidence in a criminal trial. They do not
establish the reliability of polygraphs as trial evidence, and they do
not invalidate reliability as a valid concern supporting Rule 707's
categorical ban." United States v. Scheffer (1998) 523 U.S. 303,
118 S.Ct 1261, 1266 n. 8 In other words, all Dr Holt's and the People's
talk about the "worldwide acceptance" of STR and D1S80
"loci" for CODIS or other law enforcement or investigative
purposes is simply beside the point.
5. A Particular DNA Typing Method Is Not Considered Generally
Accepted Unless Certain Essentials Of Correct Scientific Procedure Are
Met: Developmental and Internal Validation, Publication and Peer Review,
Accreditation, Proficiency Testing, Contamination Controls, and Regular
Audits
How do particular methods become generally accepted by the relevant
scientific community? One overriding principle put forth by the
community itself is that "(f)orensic DNA anlysis should be governed
by the highest standards of scientific rigor in analysis and
interpretation." NRC 1 at 52; see also, NRC 2 at75 ("It is
important that forensic laboratories use strict quality-control
standards to minimize the risk of error.") These strict standards
were first widely publicized in the 1992 NRC Report so heavily relied
upon by the Supreme Court in Venegas. The 1992 NRC Report takes no
position on the validity or scientific acceptability of any particular
forensic DNA test, but sets forth a set of requirements that it regards
as "essential" for assuring the reliability of any forensic
DNA test. The NRC II follows the same approach and specifically
recommends that "(l)aboratories should adhere to high quality
standards (such as those defined by TWGDAM and the DNA Advisory
Board)..." Id at 88 Hence, one way to judge whether or not a new
technique would be generally accepted in the scientific community (if
better known) is to determine whether it meets the "essential"
requirements established by the NRC Reports, the TWGDAM Guidelines, and
the DAB Advisory Board Standards.
The NRC Reports, TWGDAM Guidelines, and DAB Standards establish what is
scientifically acceptable procedure for conducting PCR-Based or other
DNA testing. Compliance with these and other guidelines, including the
lab's own testing protocols, is accordingly a prerequisite to
admissibility of DNA evidence in this case.
The court in People v. Barney (1992) 8 Cal. App. 4th 798, 812-813
specifically addresses this question:
The NRC Report concludes there is indeed a need for standardization of
laboratory procedures and proficiency testing (as well as appropriate
accreditation of laboratories) to ensure the quality of DNA laboratory
analysis. But the absence of such safeguards does not mean that DNA
analysis is not generally accepted....Rather, the absence of these
safeguards goes to the question whether a laboratory has complied with
generally accepted standards in a given case, or, as stated in terms,
whether the prosecutor has shown that "correct scientific
procedures were utilized in the particular case. (Emphasis added)
See also, People v. Venegas (1998) 18 Cal. 4th at 53,( "lack of
compliance by the FBI with procedures recommended in 1992 by the
National Research Council (NRC) for determining the statistical
probability of a random match" renders DNA evidence inadmissible);
United States v. Beasley (8th Cir. 1996) 102 F.3d 1440,1448( " In
every case, of course, the reliability of the proffered test results may
be challenged by showing that a scientifically sound methodology has
been undercut by sloppy handling of the samples, failure to properly
train those performing the testing, failure to follow the appropriate
protocols, and the like."); State v. Jackson (Neb. 1998) 255 Neb.
68, 582 N. W. 2d 317, 325(the results of an unspecified STR procedure
should not have been admitted absent a foundation that the lab had
followed its own testing protocols). The California Legislature also
recognizes the importence of these guidelines. See, Penal Code Section
277(d)("Nothing in this section precludes laboratories meeting
Technical Working Group On DNA Analysis Methods (TWGDAM) guidelines or
standards promulgated by the DNA Advisory Board...from performing
forensic identification analysis, including DNA profiling..."
One crucial requirement established by all of these guidelines and
standards is that forensic DNA tests must be developmentally validated
through extensive empirical studies. See,DAB Standard 8.1( "The
laboratory shall use validated methods and procedures for forensic
casework analysis....Novel forensic DNA methodologies shall undergo
developmental validation to ensure the accuracy, precision and
reproducability of the procedure."); TWGDAM Guidelines 4.1.2
(" Validation studies must have been conducted by the DNA
laboratory or scientific community prior to the adoption of a procedure
by the DNA laboratory.") According to NRC I, developmental
validation is necessary to establish an objective and quantitative
procedure for identifying the pattern of a sample and for declaring a
match, to identify possible artifacts that could lead to incorrect
interpretation if not recognized,to understand the limits of each DNA
typing method for sensitivity to quantity, mixture,and contamination, to
establish the statistical frequency of the genetic markers and the error
rate of the procedure.NRC I at 52-55
TWGDAM Guideline 4.1.5.1 sets forth the minimal requirements which must
be meet in order to satisfy the requirement of developmental validation,
including studies in the following areas: standard specimen studies,
consistency studies, population studies, reproducability studies, mixed
specimen studies, environmental studies, matrix studies, nonprobative
evidence studies, nonhuman studies, minimum sample studies, and on-site
evaluation studies. DAB Standard 8.1 contains a similar listing.
The 1992 NRC Report also emphasizes that the key validating studies must
not only be done but published and peer reviewed before a laboratory can
claim that its methods are generally accepted.
If a new DNA typing method (or a substantial variation on an existing
one) is to be used in court, publication and scientific scrutiny are
very important. Extensive empirical characterization must be undertaken.
Results must be published in appropriate scientific journals.
Publication is the mechanism that initiates the process of scientific
confirmation and eventual acceptance or rejection of a method.
NRC 1 at 56; see also, TWGDAM Guideline 4.1.5.12 ("It is
essential that the results of the developmental validation studies be
shared as soon as possible with the scientific communiy through
presentations at scientific/professional meetings. It is imperative that
details of these studies be available for peer review through timely
publications in scientific journals."). The existence of an
"extensive literature in peer-reviewed journals" was key to
the Supreme Court's recent ruling in People v. Soto, 21 Cal. 4th at 540,
that population substructure does not impact significantly upon DNA RFLP
population frequency estimates.
A second fundamental requirement is internal validation, which requires
that the laboratory gain "a solid base of experience in forensic
application" before it uses a new DNA typing method.NRC I at 55;
DAB Standard 8.1.3("Internal validation shall be performed and
documented by the laboratory.");TWGDAM Guideline 4.5("Prior to
implementing a new DNA analysis procedure or an existing DNA
procedure...the forensic laboratory must first demonstrate the
reliability of the procedure in-house.")
According to NRC I, internal validation involves five separate steps:
1. familiarity with the system using fresh samples (i.e., fresh blood);
2. test marker "survival" in dried stains (i.e., blood);
3. test system on simulated samples that have been exposed to various
environmental conditions;
4. establish "basic competence" in the use of the system by
blind trials;
5. test the system on nonprobative evidence samples in which the origin
is known, to check reliability.
The NRC I recommends that when a DNA technique is initially
developed, all five steps should be carefully followed. Id p. 55 It
further recommends that as laboratories adopt a particular technique,
"it will not always be necessary for them to repeat all the steps,
but they must demonstrate familiarity and competence by following steps
1, 4, and 5." Id.
The 1995 TWGDAM Guidelines are even more rigorous. Guideline 4.5
provides that internal validation "must" include the
following:
4.5.1 The method must be tested using known samples.
4.5.2 If a modification which materially effects the results of an
analysis has been made to an analytical procedure, the modified
procedure must be compared to the original using identical samples.
4.5.3 Precision (eg. Measurement of fragmented lengths) must be
determined by repetitive analysis to establish criteria for matching.
4.5.4 The laboratory must demonstrate that its procedures do not
introduce contamination which would lead to errors in typing.
4.5.5 The method must be tested using proficiency test samples. The
proficiency test may be administered internally, externally, or
collaboratively.
DAB Standard 8.1.3 also requires internal validation, and adds
requirements not found in TWGDAM. For instance, Standard 8.1.3.1
requires that the " procedure shall be tested using known and
nonprobative evidence samples", and that "(t)he laboratory
shall moniter and document the reproducibility and precision of the
procedure using human DNA controls." Standard 8.1.3.3 adds the
requirement that "(b)efore the introduction of a procedure into
forensic casework, the analyst or examination team shall successfully
complete a qualifying test."
In addition to requiring both developmental and internal validation, the
NRC Reports, the TWGDAM Guidelines, and the DAB Standards recommend
other measures to ensure a mininum level of quality assurance. NRC I
states that " courts should require that a proponent of DNA typing
evidence have appropriate accreditation--including documentation of
external, blind proficiency testing (as well as other accreditation that
might be mandated by government or come to be generally accepted in the
profession)--for its evidence to be admissible." NRC I at 106-107
The Report explains that
There is strong legal foundation for such a position. As a number of
courts have correctly recognized, the admissibility of scientific
evidence depends not just on a technology's being sound in principle,
but on the testing laboratory's having applied it in the case at hand
according to generally accepted standards. Courts should view the
absence of appropriate accreditation as constituting a prima facia case
that the laboratory has not complied with generlly accepted standards.
Id. at 107
See also, NRC II at 88 ("Laboratories should...make every effort to
be accredited for DNA work (by such organizations as ASCLD-LAB).";
DAB Standards, Preface("Throughout its deliberation concerning
these quality standards, the DNA Advisory Board recognized the need for
a mechanism to ensure compliance with these standards. An underlying
premise for these discussions was that accreditation would be required
to demonstrate compliance with the standards and therefore assure
quality control and a quality program. Accordingly, the Board recommends
that forensic laboratories performing DNA analysis seek such
accreditation with all deliberate speed.").
The California Legislature has embraced the accreditation requirement
wholeheartedly in the DNA and Forensic Identification Data Base And Data
Bank Act of 1998. Penal Code Section 297(a) provides that "(t)he
laboratories of the Department of Justice that are accredited by the
American Society of Crime Lab Directors Laboratory Accreditation Board (ASCLD/LAB),
and any crime designated by the Department of Justice that is accredited
by the ASCLD/LAB, are authorized to analyze crime scene samples..."
Penal Code Section 297(b) provides that "(a) biological sample
obtained from a suspect in a criminal investigation for the commission
of any crime may be analyzed for forensic identification profiles,
including DNA profiles, by the DNA Laboratory of the Department of
Justice, which is accredited by the ASCLD/LAB, or any crime laboratory
designated by the Department of Justice that is accredited by the ASCLD/LAB.
Penal Code Section 297(c) states explicitly that "(a)ll
laboratories, including the Department of Justice DNA laboratories,
contributing DNA profiles for inclusion in California's DNA databank
shall be accredited by the ASCLD/LAB.Additionally, each laboratory shall
submit to the Department of Justice for review the annual report
required by the ASCLAD/LAB which document's the laboratory's adherence
to ASCLAD/LAB standards."
As indicated, the NRC I also recommends the requirement of
"rigorous external proficiency testing via blind trials" as a
prerequisite to admissibility of DNA evidence. In fact, NRC I emphasizes
this requirement above all others:
Most important, there is no substitute for rigorous proficiency testing
via blind trials. Such proficiency testing constitutes scientific
confirmation that a laboratory's implementation of a method is valid not
only in theory, but also in practice. No laboratory should let its
results with a new DNA typing method be used in court, unless it has
undergone such proficiency testing via blind trials.
NRC 1 at 55.
The NRC II and TWGDAM Guidelines also indicate the importance of
regular proficiency testing, including blind proficiency testing. See,
NRC II at 88 (" Regular proficiency tests, both within a laboratory
and by external examiners, are one of the best ways of ensuring high
standards. To the extent that it is feasible, some of the tests should
be blind."); TWGDAM Guideline 9.2("It is highly desirable that
the DNA laboratory participate in a blind proficiency test program, and
every effort should be made to implement such a program.") The DAB
Standards are somewhat less emphatic on the requirement of blind trials,
but still it is clear that rigorous external proficiency testing is a
prerequisite to reliable DNA testing under the Standards. See,Standard
13.1(Examiners...who are actively engaged in DNA analysis shall undergo,
at regular intervals of not to exceed 180 days, external proficiency
testing in accordance with these standards."
Another requirement established in all the reports, guidelines, and
standards is the absolute necessity of controlling for the problem of
contamination in PCR testing. Forensic DNA testing based on the PCR
technique is very different than RFLP testing in this regard. There is
universal recognition in the scientific community that special and
peculiarly difficult problems must be resolved if a reliable transfer of
the extremely sensitive PCR technology from its present use in research
laboratories and clinical (medical) diagnostic laboratories to the
forensic arena is to be accomplished.
One point that emerges immediately in considering the general acceptance
of PCR based forensic testing is that the scientific community has
unequivocally warned about special dangers of sample contamination.
These dangers embrace both the methods used for the initial collection,
preservation, and handling of samples by crime scene technicians as well
as the methods used for processing the samples in the laboratory. The
NRC 1 Report expressed its most "serious concern" about the
problem of "contamination of evidence samples with other human
DNA," and warns that " (e)ven the simple act of flipping the
top of a plastic tube might aerosolize enough DNA to pose a
problem." NRC 1 at 65 The Report goes on to underscore that
reliable methods must be developed to prevent and control contamination
problems unique to forensic PCR based testing in both the sample
gathering and laboratory testing processes. NRC 1 at 66. The NRC II
likewise warns that "any procedure that uses PCR methodology is
susceptable to error by contamination" and and that "(i)f the
contaminating DNA is present at a level comparable to the target DNA,
its amplification can confound the interpretation of typing results,
possibly leading to an erroneous conclusion." NRC II at 23. The NRC
II is also emphatic that " the integrity of the chain of custody is
of paramount importance" , and that "(t)his means meticulous
care, attention to detail, and through documentation of every step of
the process, from collecting the evidence material to the final
laboratory report." Id. at 25;See also, TWGDAM Guideline
4.5.4("The laboratory must demonstrate that its procedures do not
introduce contamination which would lead to errors in typing.");
DAB Standard 6.1("The laboratory shall have a facility that is
deigned to provide adequate security and minimize contamintion."),
Standard 7.1("The laboratory shall have and follow a documented
evidence control system to ensure the integrity of physical evidence.
This system shall ensure that...the laboratory follows documented
procedures that minimize loss, contamination, and/or deleterious change
of evidence.")
Therefore, it must be emphasized that a Kelly inquiry concerning PCR
based testing must embrace the methods used by the crime scene
investigtors and laboratories for identifying, correcting, and
preventing contamination in the collection, preservation, and handling
of crime scene samples. In addition, for reasons already stated, a Kelly
inquiry must separately review the general acceptance of each PCR-based
technique, and its application by the Crime Lab in this case, whether
one chooses to characterize the inquiry as a Prong 1 or Prong 3
question.
Finally, the NRC Reports, the TWGDAM Guidelines, and the DAB Standards
stress the need for regular external audits and the need to take
corrective action in response to any deficiencies uncovered in the audit
process.See NRC I at 106(recommending periodic on-site inspections of
accredited labs); NRC II at 78-80(" Proficiency-testing and audits
are key assessment mechanisms in any program for critical
self-evaluation of laboratory performance....Regular audits of
laboratory operations complement proficiency-testing i the monitoring of
general laboratory performance."); TWGDAM Guideline 10("
Audits are an important aspect of th QA program....Audits or inspections
should be conducted at least once every 2 years by individuals separate
from and independent of the DNA testing laboratory...Records of each
inspection should be maintained and should include...remedial action
taken to resolve existing problems..."); DAB Standard 14 (The
laboratory shall establish and follow procedures for corrective action),
Standard 15(The laboratory shall conduct audits annually and once every
two years, a second agency shall participate in the annual audit)
These emphatic pronouncements, from distinguished scientific panels,
following many years of study of the issues, hold an important lesson.
The court should not take seriously the claim that a new DNA typing
method is "generally accepted" within the meaning of Kelly if
the method has not been developmentally and internally validated by
extensive publication and peer review, and if the particular lab does
not meet the accreditation, proficiency testing, contamination control,
and audit requirements set forth in nationally recognized standards for
the quality assurance of forensic DNA testing laboratories.
II.
DEFENDANT ROBERT NAWI IS ENTITLED TO CHALLENGE THE BLOODTYPING RESULTS
IN THIS CASE UNDER KELLY-FRYE BECAUSE ISSUES RAISED IN THIS OPPOSITION
HAVE NOT BEEN RESOLVED BY ANY APPELLATE COURT IN CALIFORNIA, NOR IN ANY
CASE PENDING REVIEW BEFORE THE CALIFORNIA SUPREME COURT
Citable case law does not answer all of the questions that arise when
deciding the admissibility of DNA typing evidence. Some issues regarding
the admissibility of a small number of DNA typing tests have been
addressed by the California Supreme Court and by California Courts of
Appeal. However, as will be discussed in detail below, neither citable
California case law nor the cases that for many years were pending
review before the California Supreme Court contain issues that
definitively decide the issue of admissibility of PCR-based DNA testing.
A. THE ISSUES RAISED IN THIS MOTION ARE NOT INCLUDED IN ANY CASE
CURRENTLY PENDING BEFORE THE CALIFORNIA SUPREME COURT
Criminal lawyers in California, defense and prosecution alike, have
been anxiously awaiting decisions by the California Supreme Court in
nine cases, all of which raise Kelly-Frye issues regarding the
admissibility of either RFLP or PCR-based DNA testing. On May 11, 1998
the California Supreme Court issued a much-anticipated decision in
People v. Venegas, supra. On August 23, 1999, the Court decided People
v. Soto, supra. Although Venegas and Soto involved a Kelly-Frye
challenge of RFLP technology, the holding in these cases provides the
California Supreme Court's most recent interpretation of the parameters
of Kelly-Frye. Seven other cases were for many years on hold at the
Supreme Court, pending action in light of the decisions in Venegas and
Soto . Very recently, on October 27, 1999, the Court dismissed review in
all seven cases , thus leaving the law as it relates to PCR-based
testing in some doubt. The defense recognizes that these seven cases are
not citable authority. However, the following discussion is presented
for the purpose of distinguishing the issues raised by defendant Robert
Nawi in this motion from those issues for which resolution by the
Supreme Court was anticipated.
On March 16, 1995 the California Supreme Court granted review of the
Fifth District appellate decision in People v. Venegas (1995) 31
Cal.App.4th 234. Sergio Venegas was charged with rape and other
offenses. The FBI conducted RFLP testing on samples taken from the
defendant, semen stains on the bedspread on which the rape occurred and
vaginal swabs taken from the victim. Venegas, supra, at 234. The FBI
concluded that for three of the four tests conducted, the defendant's
DNA matched the DNA in the crime scene samples. Id. A fourth test was
inconclusive. Applying statistical methods, the FBI concluded that the
likelihood of another individual being the source of the DNA found at
the scene was one in 65,000 of the general population and one in 30,000
of the southwestern Hispanic population. Id.
In the court of appeal, the defendant prevailed on a "prong
one" Kelly claim that the trial court had erred in finding general
acceptance in the scientific community for the RFLP method and the
statistical methods used in conjunction with RFLP. In addition, the
defendant prevailed on a "prong three" that the FBI failed to
use correct scientific procedures in this case in violation of Kelly,
supra, 17 Cal.3d at 30.
The Supreme Court only partially agreed with the Court of Appeals.
Regarding Mr. Venegas's prong one challenge to the general acceptance of
RFLP, the Supreme Court followed the rulings in People v. Barney and
People v. Axell and held that the RFLP methodologies used by the FBI
were generally accepted in the scientific community.
The Supreme Court also ruled on Mr. Venegas's prong three challenge.
After a discussion of the nature of the third prong of Kelly, the
Supreme Court found that the FBI failed to follow the recommendations of
the 1992 NRC Report in the statistical calculation step of their
analysis. Thus, the Supreme Court upheld the Court of Appeal's reversal
of Mr. Venegas's conviction, and remanded the case. In two footnotes,
the court expressly left open any Kelly issues regarding PCR testing 18
Cal.4th at 58, n.5 and the issue of whether the statistical calculations
must account for laboratory error, 18 Cal.4th at 95, n.42.
On the same day that the Supreme Court granted review in Venegas, that
Court granted review to People v. Soto (1995) 30 Cal.App.4th 340. Frank
Soto was convicted of one count of attempted rape. Using the RFLP
method, the Orange County Sheriff's Department laboratory compared DNA
from defendant's blood with DNA taken from semen found at the crime
scene and concluded that, using the "fixed bin"statistical
method the likelihood that someone other than the defendant was the
source of the semen stain was one in 189 million Hispanics and one in 38
million Caucasians.
On appeal the defendant challenged the use of the product rule, a step
in the statistical analysis where the probability of each locus of
specific lengths appearing in the general population are multiplied
together. The product results in extremely high probabilities that the
defendant is the source of the sample in question. The Court of Appeal
found that although there were scientists who did not agree that the
product rule was not appropriately used in RFLP statistical
calculations, the product rule is generally accepted in the relevant
scientific community.
In People v. Soto (1999) 21 Cal. 4th 512, the Supreme Court upheld the
court of appeals decision, finding that " (i)t is clear from the
evidence in the record, the clear weight of judicial authority, and the
published scientific commentary, that the unmodified product rule, as
used in the DNA forensic analysis of this case, has gained general
acceptance in the relevant scientific community and therefore meets the
Kelly standard for admississibility." Id at 541 However, the
Court's holding was founded in the facts adduced at the hearing that an
"extensive literature in peer-reviewed journals" has
accumulated in support of the conclusion that population substructuring
does not impact significantly upon RFLP population frequency estimates.
Id. at 540-541. The Court cited a worldwide study of RFLP frequency data
by the FBI; the publication of the NRC 2 in 1996 which concluded that if
the tests are properly conducted the unmodified product rule can be
used; and the proliferation of peer reviewed literature supporting the
conclusion that for RFLP tests, the use of the product rule is
appropriate. The court concluded: "It is clear . . . that the
unmodified product rule, as used in the DNA forensic analysis in this
case, has gained general acceptance." Id. at 541 [emphasis added].
The Court went on to limit its holding to RFLP cases in which the
probabilities of a random match are very rare--one in the multimillions
or billions: " We have no occassion in this case to consider
whether substructuring could be a cause of material variation in much
higher frequencies, e.g., one in several hundred." Id. at 541.
Thus, nothing in the Soto decision impacts on the issues in this case.
First, it was an RFLP case in which the Court attached the
"greatest significance" to NRC II. 21 Cal. 4th at 539. The NRC
II stresses the tentative nature of our current state of knowledge with
respest to PCR-based systems, as well as a fundamental distinction
between RFLP (VNTR) loci and PCR loci:
PCR-based systems have several advantages, the most important being that
they can be used when source material is sparce or degraded and a second
being that there need not be uncertainties of measurement. But there are
also disadvantages. VNTRs have many allelles, none of which is at a high
frequency. Presumably, the high mutation rate accounts for that and for
the small differences in frequencies among subgroups. The VNTR's used
for forensics also occur at loci that have no function and therefore are
probably not affected by natural selection. Some of the loci used in PCR-based
systems have only a small number of alleles, and the loci are at
functional genes, which means that there is less assurance for HW (Hardy
Weinberg equilibrium) and LE (linkage equilibrium). Many more loci are
required to produce the same probability levels than are required for
VNTRs.
Yet, the statistical uncertainties with VNTRs...make it desirable to
bring new loci into the system. The extensive activity in mapping genes
is leading to the rapid discovery of many more possible markers, some of
which are expected to have the kinds of properties that are desirable
for forensic use: high mutation rate, multiple alleles, lack of function
(which increases the possibility of neutrality), speed of analysis, low
cost, and unambigious identification of alleles. We encourage the
development and validation of such systems.
NRC II at 120-121
NRC II also explicitly recognizes that "(t)here has not been the
extensive sampling of subpopulations and geographic areas for PCR-based
systems that has been done with VNTR's." Id. at 35. Based on this
lack of research and based further on emerging data which indicates both
that PCR loci are at functional genes, and that population
substructuring does impact significantly upon PCR population frequency
estimates, prominent population geneticists testified in Bokin that the
unmodified product rule is inappropriate for PCR-based tests. Moreover,
NRC II itself clearly proposes modifications of the product rule for
RFLP and PCR-based systems, such as the "2p rule", the use of
a (theta) factor, and the use of confidence intervals. Id. at 119-120,
146. Nothing in Soto abrogates the use of these modifications of the
frequency calculations using the product rule.
Second, Soto solely involves statistics and the use of the unmodified
versus the modified product rule to calculate RFLP match frequencies in
an unmixed sample, and not questions of methodology. See, People v. Soto,
21 Cal. 4th at 518 (" He focused his challenge, not on the
threshold RFLP analysis or on the comparison of his DNA with that found
on the bedspread, but on the second stage of the analysis: the
population frequency determination.") Here, Mr. Nawi's challenge is
focused on all aspects of PCR analysis: the threshold DQ alpha,
Polymarker, Green I, and Profiler Plus analysis, the comparison analysis
for each of these systems, and the population frequency determination on
a mixed sample. Soto does not address these questions.
Also on March 15, 1995, the Supreme Court agreed to review the Fourth
Appellate District's decision in People v. Wilds (1995) 31 Cal.App.4th
636. Defendant Wilds was convicted of robbery, two counts of forcible
oral copulation and rape of one victim; robbery, three counts of
forcible oral copulation and rape of another victim; and burglary of a
third victim. From DNA testing conducted by the RFLP method, the police
concluded that the sperm cells found in the victims were
"identical" to the DNA taken from the defendant's blood.
Experts testified that the probability of someone other than the
defendant being the source of the sperm as one in 4.5 million.
On appeal Wilds argued that both the RFLP methodology and the
statistical calculations were not generally accepted in the scientific
community, and that the trial court erred in so finding. The First
Appellate District did not agree. Citing People v. Axell (1991) 235
Cal.App.3d 836; People v. Barney (1992) 8 Cal.App.4th 798; People v.
Leahy (1992) 8 Cal. 4th 587, 612; and People v. Soto (1994) 30
Cal.App.4th 340, the court concluded that RFLP typing and the related
statistical methods are admissible under Kelly. As indicated above, on
October 27, 1999, the Court dismissed the petition for review under Cal.
Rules of Court 29.4(c). That Rule provides that the Supreme Court may
dismiss review of a cause as "improvidently granted", and that
upon the dismissal, the opinion of the Court of Appeal shall become
final but shall remain unpublished unless the Supreme Court expressly
orders otherwise. No such publication order was entered in Wilds.
One month after Venegas, Soto, and Wilds the Supreme Court granted
review to another DNA bloodtyping case. In People v. Marlow (1995) 48
Cal.App.4th 1440, review granted, April 25, 1995 (No. S046996) defendant
Gustavo appealed his convictions of, inter alia, two counts of murder,
rape, and kidnaping. Marlow argued that not only was RFLP not generally
accepted in the scientific community, but also the lab that conducted
the RFLP in his particular case, failed to follow proper procedures in
conducting their analysis. The Sixth District Court of Appeal found
that, although there was some disagreement, RFLP is generally accepted
in the scientific community, and that the lab used proper procedures. On
October 27, 1999, the Court dismissed the petition for review under Cal.
Rules of Court 29.4(c), with no order that the Court of Appeal opinion
be published.
Venegas, Soto, Wilds and Marlow each raised essentially the same issues:
whether there is general acceptance in the relevant scientific community
of RFLP and the statistical methods used in conjunction with RFLP, and,
in the case of Venegas and Marlow, whether proper procedures were used
in their particular cases. The next case granted review by the Supreme
Court raised a new issue. In People v. Amundson, (1995) 34 Cal.App.4th
1151, review granted August 10, 1995 (No. S047242) Wayne Amundson
challenged, inter alia, his conviction of murder with a special
circumstance allegation that the murder was committed during a rape.
Cellmark diagnostics conducted RFLP analysis. The Serological Research
Institute ("SERI"), using PCR to amplify the DNA samples,
compared the alleles of DQ alpha locus in samples taken from the
defendant and samples taken from the crime scene. In addition to
challenging the statistical calculation step of RFLP, the defendant
argued that amplifying the DQ alpha locus using the PCR technique for
the purposes of identifying an individual as the source of DNA was not
generally accepted in the scientific community.
The Court of Appeals for the Fourth Appellate District did not agree.
Basing its conclusion on the trial record, case law from other
jurisdictions, and scientific literature, the court found that "PCR
testing" is generally accepted in the scientific community.
Amundson, supra, at 803. The Court also held that the statistical
methods used in RFLP protocols were also admissible under Kelly-Frye. On
October 27, 1999, the Supreme Court dismissed the petition for review
under Cal. Rules of Court 29.4(c), with no order that the Court of
Appeal opinion be published.
Next in the line of cases on hold for many years was People v. Burks
(1995) 41 Cal.App.4th 652, review granted Nov. 16, 1995 (No. S048916).
Melvin Burks was convicted of, inter alia, forcible rape, and two counts
of sexual battery by restraint. Using RFLP methods, a lab hired by the
prosecution compared semen found at the scene of both attacks with a DNA
sample taken from the defendant, and declared a match. Id.
On appeal, the defendant argued that the DNA evidence was inadmissible
because the use of the modified ceiling principle in statistical
calculations is not generally accepted in the scientific community. The
court found that because the " ‘modified ceiling' method employed
here in the statistical calculation step of RFLP analysis of DNA is more
favorable to Burks than the ‘unmodified product rule' rejected in
Barney and is generally accepted withing the scientific community …
RFLP analysis using the more conservative ‘modified ceiling' approach
satisfies Kelly evidentiary requirements." On August 12, 1998, the
Supreme Court dismissed the petition for review under Cal. Rules of
Court 29.4(c), with no order that the Court of Appeal opinion be
published.
On April 25, 1996 the Supreme Court granted review to People v. Smith
(1996) 42 Cal.App.4th 204, review granted April 25, 1996 (No. S052233).
In this case defendant Frank Smith appealed his conviction of, inter
alia, two counts of forcible rape. Smith argued to the Fourth Appellate
District that the trial court erred in admitting bloodtyping evidence.
Specifically, Smith challenged the statistical methods, specifically the
product rule, used as part of the RFLP analysis by the Orange County
Crime Lab.
In rejecting Smith's argument, the Court of Appeal noted that the trial
court had found that the Orange County Crime Lab had followed correct
procedures in that case, and that the question regarding the use of the
product rule in RFLP cases had been decided in People v. Axell, supra.
The court concluded that the use of the product rule was generally
accepted in the scientific community. On October 27, 1999, the Supreme
Court dismissed the petition for review under Cal. Rules of Court
29.4(c), with no order that the Court of Appeal opinion be published.
The Supreme Court also granted review in People v. Givens (1997) 53
Cal.App.4th 554 review granted May 28, 1997 (No. S060563). Harold Givens
was convicted of rape, forcible oral copulation, and burglary. The
convictions arose out of attacks on four women. For each victim, PCR and
RFLP testing was conducted on samples taken from vaginal swabs. PCR
testing was conducted by Alan Keel, the criminalist in the case at bar;
RFLP analysis was conducted by Cellmark, a private laboratory.
The focus of Givens's claims to the Court of Appeal and of that Court's
decision was the use of the unmodified product rule calculations. Givens
argued that the trial court improperly admitted the results of the DNA
tests because the unmodified product rule was not generally accepted in
the scientific community. Although the general acceptance of PCR was at
issue in the Kelly-Frye hearing held by the trial court, this claim was
not addressed by the Court of Appeal.
The First District disagreed with Givens, "[b]ased on the quality
and the quantity of evidence produced in the instant case" Givens,
supra, 53 Cal.App.4th at 668, and held that the DNA typing evidence was
properly admitted. As with the other cases, on October 27, 1999, the
Supreme Court dismissed the petition for review under Cal. Rules of
Court 29.4(c), with no order that the Court of Appeal opinion be
published.
Finally, and most recently, the Court granted review in People v.
Daniels (1998) 62 Cal. App. 4th 1529, review granted June 17, 1998.Mr.
Daniels was convicted of rape based on a blood sample that was analyzed
by RFLP, DQ alpha, and D1S80 testing. The PCR testing was done by the
Santa Clara County Crime Lab. At trial, the defendant did not object to
the PCR test results, but did object to the RFLP statistical
calculations performed by the F.B.I. On appeal, the defendant again
raised the objection to the use of the product rule for RFLP testing,
and also claimed that his trial counsel was ineffective for failing to
object to the use of the product rule for the PCR testing. In a brief,
two paragraph discussion the court rejected both arguments. With respevt
to PCR and the possible applicability of Barney, the court stated that
"there is no basis for extrapolating its conclusions about RFLP
match frequencies to PCR procedures, a completly different method of DNA
analysis." Id. As with the other cases, on October 27, 1999, the
Supreme Court dismissed the petition for review under Cal. Rules of
Court 29.4(c), with no order that the Court of Appeal opinion be
published.
Equally important as determining which issues have been in front of the
Supreme Court is identifying which have not been in front of that Court.
First, Amundson and, to a limited degree, Givens raised only one issue
relating to PCR-based tests. In Amundson SERI amplified only the DQ
Alpha locus. In addition, as mentioned above, the general acceptance of
PCR-based testing was not raised on appeal in Givens. Thus, the only
issue that could have been resolved if the Supreme Court had not
dismissed review in Amundson and Givens was whether using PCR-based
testing to amplify the DQ Alpha locus is generally accepted in the
scientific community. In any event, the Court's action in these cases
has left even this narrow question unresolved by the Supreme Court. This
opposition includes challenges to the DQA1+PM, D1S80, Green I and
Profiler-Plus tests employed by the Crime Lab in Mr. Nawi's
case.Moreover, none of the cases before the Court involve the use of
twelve year old fingernail samples allegedly containing complicated
mixtures of several people, which raises issues unique to this case.
The cases recently before the Court also do not consider the issue of
whether the databases used in determining the significance of
similarities between PCR samples are reliable and generally accepted, or
whether an error rate or other statistical modifications are necessary
for PCR-based systems.. Although other states have considered some of
these issues, see, e.g., State v. Carter (Neb. 1994) 524 N.W.2d 763,
overruled on other grounds, 571 N.W.2d 276, 293, this issue is not
presently before the Supreme Court nor, as will be detailed below, has
it been decided by a California Court of Appeal.
Finally, neither Venegas, Soto, nor any of the other cases raised the
issue of whether the San Francisco Police Crime lab uses proper
procedures in conducting PCR-based testing, although Venegas does make
clear that a hearing must be conducted on this issue in this case.
B. ISSUES RAISED IN THIS MOTION HAVE NOT BEEN RESOLVED BY ANY
CALIFORNIA COURT OF APPEAL.
In 1998, the First District issued a written opinion on the
admissibility of DNA testing using an earlier version of Roche Molecular
Systems' DQ Alpha kit and an earlier version of the Polymarker kit.
People v. Wright (1998) 62 Cal.App.4th 31. The Wright court relied
solely on the holding in People v. Morganti, infra, in holding that both
the DQAlpha and PM kits were admissible under Kelly-Frye, stating
"[t]he PCR matching technique, at the very least, has certainly
acquired general acceptance in the scientific community." Id. at
1832. For a number of reasons, Wright cannot be considered precedent or
controlling on the issue of whether the Crime Lab's PCR based DNA
testing methods meet Kelly-Frye standards.
First, the Wright court misread the holding of Morganti. Wright states
that "(f)ollowing Morganti, we agree with the trial court's ruling
here. The PCR method has obviously acquired general acceptance in the
scientific community." Id. at 1834. Morganti did not hold that the
PCR method was generally accepted; Morganti only considered the
acceptance of an early version of Cetus's original DQAlpha kit, 43 Cal.
App. 4th 643, 662 n. 9, which as laid out above, and which will be
developed further infra, differs from the DQ Alpha +PM kit, the Green I
kit, and the Profiler Plus kit employed in this case. In addition, when
DQ Alpha is combined into a single kit with the five polymarker loci and
then used in conjunction with two STR kits to come up with a combined
probability, issues not relevant to DQ arise that could not have been
addressed in Morganti, specifically the use of population genetics and
statistical analysis of multiple genes, and the effects on reliability
of combining a number of loci into a single DQ Alpha + Polymarker kit.
As the NRC II recognized following the decision in Morganti,
the Polymarker system "was only beginning to be widely used"
in 1996. NRC II at 72. As use of this system has developed, more recent
scientific literature calls into serious question whether the community
of molecular biologists accept as reliable the DQ alpha + Polymarker kit
used in this case. For example, one recent article regarding the
combined DQ Alpha +PM kit found that attempts at typing samples using
this multiplex kit resulted in failed hybridization and allelic dropout.
Donald E. Riley, Compromised Optimal Hybridization Temperatures May
Explain PM Plus DAQ1 Anomalies, 44 Biochemistry and Molecular Biology
International 59, 63 (1998). Dr. Riley testified in the Bokin hearing
and explained extensively his study and the basis for his expert opinion
that the DQ alpha + PM kit was not generally accepted as reliable by the
community of molecular biologists. Dr. Riley's article was not and could
not have been considered in either Wright or Morganti, because it was
not published until after the decisions in those cases. As Soto
emphasizes, "(i)n the context of rapidly changing technology, every
effort should be made to base (decision) on the very latest scientific
opinions, including those published during the appellate phase of the
case." 21 Cal. 4th at 540 n. 31
What the Wright court failed to consider is that the while PCR is a
method used in DNA testing, it is only one step of many used to
determine a genotype and to calculate population frequencies.See, United
States v. Hicks (9th Cir. 1996) 103 F. 3d 837,844 [ "The PCR method
itself is not a genetic test; it is a mere amplification
technique."]. An analogy to the Wright court's myopic focus on PCR
would be a court holding that using ink to take fingerprints was
generally accepted but failed to consider the further steps taken to
analyze the print. It was not until our Supreme Court issued its ruling
in Venegas, decided three months after Wright, that it became clear that
California courts were required to scrutinize each step of DNA analysis
under Kelly. Venegas, and not Wright, sets forth the requirement of
addressing DNA evidence at each stage of the process.
The NRC in its first report recognized the distinction between different
tests using the common step of PCR, recommending that in "PCR-based
typing, the committee makes a number of technical recommendations,
including recommendations for thorough characterization of each PCR
assay for definition of the range of conditions under which it will
perform reliably." NRC 1, at 72-73 [emphasis added]. If all tests
that used PCR as a basis could be lumped together under the general
title "PCR" then the recommendations of the NRC to
characterize the different assays used and the validation studies done
on each new marker or locus, see infra, would be meaningless.
Therefore, the holding of Wright is not controlling in this case because
it incorrectly lumped together all tests using PCR technology in its
interpretation of Morganti. As explained above, the DQ Alpha + PM kit
differs from DQ Alpha alone in that it analyzes different genes or
makers and therefore use different primers in the amplification step and
different reverse dot blot strips in the hybridization step, and the use
of multiple markers also complicates the statistical calculations. Also,
as demonstrated below, the D1S80 kit, the Green I kit, and the Profiler
Plus kit raise unique reliability problems which were not and could not
have been considered in either Wright or Morganti.
Second, Wright and Morganti are not controlling precedent because new
evidence of a controversy is now available and must be considered under
the reasoning of Kelly. Although Kelly states that once a trial court's
decision to admit new scientific evidence is affirmed on appeal and the
"precedent so established may control subsequent trials," the
court also recognized that the decision will only be precedent
"until new evidence is presented reflecting a change in the
attitude of the scientific community." 17 Cal.3d at 32 [emphasis
added].
The holding in Morganti, supra, was limited to the admissibility of DQ
PCR testing under Kelly-Frye. Id. at 663. The PCR analysis conducted and
reviewed by the Morganti court involved only the DQ gene. Id. at 662.
The court described the procedure in the following manner:
In the forensic setting, PCR analysis of DQ alpha involves three general
steps. First, DNA is extracted from the nucleus of cells present in an
unknown bloodstain. Second, the DQ alpha is replicated or amplified by a
process which involves combining the DNA with a commercially available
solution or ‘cocktail' and then subjecting the solution to a series of
controlled temperature cycles. Finally, the amplified gene is typed in
order to identify the alleles present in the amplified DNA.
Id. [emphasis added].
At the Kelly-Frye hearing, the People called two witnesses: Edward Blake
and Gary Harmor, identified by the trial court as "forensic
serologists." Id. at 664. Harmor testified that "he had not
read any publications which criticize PCR analysis of DQ alpha as flawed
or unreliable. And he was not aware of any criticism of the statistical
data he used." Id. Blake testified that the technique was accepted
as reliable in the scientific community and that he was "not aware
of any scientist who work in this field or of any publications
expressing the view that PCR analysis is unreliable." Id. at 644
n.11. Blake further testified that the NRC 1 report says that the PCR is
an "acceptable reliable process." People v. Morganti, No.
19691, Superior Court of Sonoma County, April 5, 1993 (R.T. 2935).
The only witness who challenged the assertions of Blake and Harmor was
Dr. Grunbaum, who stated that while PCR was generally accepted among
scientists, its application in the forensic arena was not generally
accepted. Id. at 664. The court rejected Dr. Grunbaum's opinion as
having little weight because Dr. Grunbaum appeared to be biased against
Harmor, he had little forensic experience and his opinions were
"unsupported by reason or lacked foundation." Id. The fatal
flaw of Morganti is that neither the trial court nor the court of appeal
was provided with a fair overview of the current underlying controversy
regarding PCR, as is done in section III of this brief. Most of this
controversy developed after the Kelly hearing in Morganti thus bringing
it within the ambit of Kelly's rule that new evidence can be considered
if there is a change in the attitude of the scientific community.
At the time of the Kelly-Frye hearing in Morganti, the Polymarker kit
was not available. The Kelly-Frye hearing was held in April 1993, a few
months prior to the introduction of the PM test kit by Perkin-Elmer.
Therefore, at the time of the hearing, the court could not have, nor did
it, consider the use of the DQ Alpha in conjunction with the use of the
PM kit. Finally, since the time of Morganti, a new kit was developed
that allows the amplification and typing of the DQ and PM markers
together. This combination required extensive new testing and changes in
the protocols used. See Ann Marie Gross & Richard Guerrieri, HLA
DQA1 and Polymarker Validations for Forensic Casework: Standard
Specimens, Reproducibility, and Mixed Specimens, Journal of Forensic
Sciences (1996) 41(6):1022. As indicated above, one recent article
regarding the combined DQ Alpha +PM kit found that attempts at typing
samples using this multiplex kit resulted in failed hybridization and
allelic dropout. Donald E. Riley, Compromised Optimal Hybridization
Temperatures May Explain PM Plus DAQ1 Anomalies, 44 Biochemistry and
Molecular Biology International 59, 63 (1998). Surveying prior studies,
Dr. Riley states that "three independent studies reported false
typings with either DQ Alpha and Polymarker. In two of these studies,
the false typings were confirmed by other techniques such as DNA
sequence data or by comparisions with nonmultiplex systems." Id. p.
64 Other recent scientific commentary regarding the general
unreliability of PCR-based testing is surveyed in section III infra.
In addition, with the use of multiple markers, new issues regarding the
use of population genetics and statistical analysis in calculating
genotype frequencies have arisen based solely on the use of more than
one gene as a marker. See, infra.
Moreover, Wright itself recognizes a variation of the "change in
scientific attitude rule" of Kelly when it observed that
"appellant also does not provide any arguments specifically
directed against the result reached in the Morganti decision." 62
Cal. App. 4th at 38. Here, however, such arguments are being raised. In
addition to the issues raised in Point III infra, it is important to
stress that both the trial court and the court of appeal in Morganti
relied on Blake's representation of the opinion of the National Research
Council and the scientific community. Review of the NRC 1 shows that
Blake provided misleading testimony to the trial court.
NRC 1, which was published in 1992, does not state that PCR or the DQ
Alpha system is reliable or generally accepted. Instead, NRC 1 provides
only technical descriptions and passing comments about the potential
uses of PCR- based testing of forensic samples. Indeed, NRC 1 identifies
areas of research that need to be pursued. "PCR provides excellent
starting material for direct DNA sequencing, and sequence analysis might
ultimately be the approach used for personal identification. But it will
require improvements in automated sequencing technology and the
generation of larger databases on sequence variability." NRC 1 at
44. The NRC went on to note that " further experience should be
gained with respect to PCR in identity testing." NRC 1 at 70.
Finally, the NRC concluded, "PCR analysis is extremely powerful in
medical technology, but it has not yet achieved full acceptance in the
forensic setting." NRC 1 at 70.
Recently, misrepresentation of the views of the National Research
Council by an expert to the court resulted in a reversal of a
defendant's convictions of first degree murder, burglary, and sexual
battery. Murray v. Florida (1997) 692 So.2d 157, 163. In Murray the
Supreme Court of Florida found that the prosecution had failed to meet
its burden to show that PCR methodologies were generally accepted
because the prosecution's witness avoided questions regarding his
testing procedures, gave "unenlightening" testimony regarding
his probability calculations and, most importantly, "affirmatively
misled the trial court as to the NRC's acceptance of PCR DNA methodology
at the time of the hearing." 62 So.2d at 163. The court explained
that
[Prosecution expert] Nippes told the court that the report of the [NRC
1] ‘was an endorsement of the forensic DNA application to forensic
science whether it was RFLP or PCR." In fact, the NRC's 1992 report
expressly withheld endorsement of PCR methodology. The committee has
explained that although the PCR method has "enourmous promise"
"it has not yet achieved full acceptance in the forensic
setting." NRC Report at 70. Specifically, in relation to the use of
PCR testing with commercial kits, one of which was used in this case,
the NRC report cautioned: "The Committee sees a potential for
introduction of unreliable kits and the musiuse of kits. The existence
of kits suggests ease of use and low chance of technical error. The
committee believes that nonexpert laboratories will run a significant
chance of error in using kits" NRC Report at 69.
692 So.2d at 160, n.5
The court went on to hold that "the paucity of information in this
record concerning the application of the PCR methodology to the DNA
evidence at issue here leads us to the conclusion that even if the trial
court had attempted to determine whether this evidence met the Frye
standard, there is no way the court could have found it
admissible." Id. at 63. This court should follow Murray and
conclude, as suggested in Wright itself, that because of the
misrepresentation of Dr. Blake, there are valid arguments specifically
directed against the result reached in Morganti.
Moreover, it should go without saying that whatever his status when the
Kelly hearing was held in Morganti in April 1993, Dr. Blake is currently
" too closely identified with the endorsement of [the technique] to
assess fairly and impartially the nature and extent of any opposing
scientific views." Kelly, supra, 17 Cal.3d at 38. Thus, when
applying the Kelly standard, the court must look to experts who are
"'impartial,' that is, not so personally invested in establishing
the technique's acceptance that he might not be objective about
disagreements within the relevant scientific community." People v.
Brown, supra, 40 Cal.3d 512, 530; accord People v. Venegas, 18 Cal. 4th
at 77 (" FBI Agent Lynch, though vigorously defending the merits of
the FBI's RFLP analytical procedures she had followed in this case, did
not purport to be qualified, as a molecular biologist or otherwise, to
testify on questions of general scientific acceptance of the validity of
those prosedures.")
Nowhere is Dr. Blake's lack of impartiality better illustrated than in
his recent intemperate and unfair personal attacks against Judge Robert
Dondero. Blake, who has admitted in several Kelly hearings that he makes
his living promoting and conducting PCR-based forensic testing, was in
1996 awarded a lucrative city contract to supervise the San Francisco
Crime Lab's DNA Program. Lashing out at the recent ruling of Superior
Court Judge Robert Dondero in People v. Bokin, which was highly critical
of the SFPD's DNA Program, Blake has called for the removal of Judge
Dondero from office on the outrageous ground that the judge is "an
arrogant, ignorant, stupid judge who is unable to get through the smoke
thrown at him by a very skilled lawyer.""Judge Rejects DNA
Test Evidence", San Francisco Examiner, May 7, 1999, p1. His
intemperence has not escaped the criticisms of his scientific collegues.
See, Comments of Professor William Thompson, People v. Bokin: An
Internet Discussion, http://CA_v_Bokin.pdf@www.scientific.org ("We
will get a lot farther through respectful exchanges...than we will by
publically declaring that those who disagree with us are ignorant,
stupid or callous."). To give credence to the views of Blake in
these circumstances would make a mockery of Kelly's prong two
requirement of impartiality. The same most certainly must be said for
the Roche Molecular Systems and Cetus employees who authored the alleged
"validation" studies upon which the Morganti court relied. 43
Cal. App. 4th at 664 665. Scientists who are affiliated with companies
that deliberately lie and distort scientific data to the Unites States
government in order to obtain lucrative patent rights cannot be
considered impartial, and their so-called scientific "studies"
cannot be allowed to pollute the waters of justice.
Finally, Wright should not be considered controlling due the manner in
which the Kelly-Frye hearing was conducted in the trial court. The
defense did not call any expert witnesses to testify at the Kelly-Frye
hearing and chose instead to call an expert to attack the DNA evidence
at trial. The court of appeal refused to apply the defense experts
testimony to the issue of Kelly-Frye, holding the decision to not call
the expert at the hearing a tactical decision by the trial attorney. 62
Cal. App. 4th at 41 Because of this tactical decision, the trial court
and the court of appeal were presented with a Kelly-Frye hearing in
which three prosecution expert witnesses testified with no opposition
from the defense. The Wright court's resolution of this issue, and its
refusal to consider the defense expert's testimony, are disapproved in
Soto, which emphasizes that, "(i)n the context of rapidly changing
technology, every effort should be made to base (decision) on the very
latest scientific opinions, including those published during the
appellate phase of the case." 21 Cal. 4th at 540 n. 31
The hearing in Wright leaves this Court in a position to consider new
evidence of a controversy that was not presented in Wright and that is
presented here. Therefore, this Court cannot find that Wright precludes
a Kelly-Frye hearing in this case. In this case, evidence that was not
presented in either Wright or Morganti regarding the use of both the Dq
+ PM, the D1S80 kit, the Green I kit , and the Profiler Plus kit, alone
and in combination, has been presented to this Court, thus requiring a
Kelly-Frye hearing.
The People have argued in McClanahan and Bokin that California case law
holds that all PCR-based tests pass Kelly-Frye, and that there is no
need for additional Kelly-Frye hearings when all that is being changed
is the gene which is being amplified and visualized. The National
Research Council and the numerous cases cited in section D. 2. Supra do
not agree. As the NRC I summarises the issue, "[T]he use of various
detection technologies for PCR products might require a pretrial hearing
about the characteristics of the detection method and its sensitivity to
artifacts. In each case, the court can properly limit inquiry to the
substantially novel aspects of the technology, focusing on whether the
method is accepted for scientific applications and whether it has been
validated for forensic identification." NRC 1 at 144-45.
The People have further argued that the NRC 2 makes quite clear that it
finds PCR technology, including the methodologies and genetic loci typed
in this case to be reliable. The People's representation of the NRC 2 is
misleading. The People's have relied on quotes from the NRC 2 involving
only general statements about "DNA typing." These quotes,
coming from a small group of primarily statisticians and lawyers, reveal
nothing about the general acceptance or reliability of the typing
methodologies used by the Crime Lab. The People have also in the past
quoted the NRC 2's description of the advantages of PCR-based tests over
RFLP testing. Again, this statement provides no insight regarding the
reliability and general acceptance of the typing methodologies used in
this case, and more importantly, the People tellingly omit the remainder
of the NRC's discussion which focuses on the disadvantages of PCR-based
testing, including that
" any procedure that uses PCR methodology is susceptable to error
by contamination." NRC II at 23 Next, the People have referred to
the reaffirmation by the NRC 2 that "the molecular technology is
thoroughly sound and that the results are highly reproducible" ,
quoting NRC 2 at 23. The NRC 1 found that the technology of PCR
amplification in general to be sound, but did not endorse genotyping of
any particular locus. See, e.g., NRC 1 at 44.
The People have cited the NRC 2 for the proposition that the polymarker,
D1S80, and STR loci have been validated for forensic use. However, as
will be discussed in more detail below, the tests used in this case have
not been validated in the manner required by NRC 1, the TWGDAM
Guidelines, or the DAB Standards. Also as will be discussed below, there
exists controversy regarding the reliability of typing these loci. The
broad statements in the NCR 2 cannot be interpreted as general
acceptance of the DQA1+PM D1S80, and STR kits used in this case,
especially since the STR kits did not even become available until after
the publication of NRC II in 1996.
Finally, the last two quotes provided by the People in other cases are
taken out of context. The first, "[w]e conclude that PCR-based
systems should be used." NRC 2 at 119, comes at the end of a
chapter on population genetics. This chapter discusses the population
biology of the loci that are studied in PCR-based systems. Specifically,
the chapter discusses issues such as independence of alleles. This
chapter does not discuss at all the reliability of amplifying particular
loci or the genotyping of particular loci. The paragraph containing this
quote ends with admonishment to use population data from PCR-based
systems conservatively. The People also have quoted another broad
statement by the NRC regarding "profiling technology", quoting
NRC 2 at 73. This vague statement comes at the end of a chapter
containing descriptions of all available DNA profiling techniques,
including RFLP, PCR-based tests, and newer tests known as mitochondrial
DNA testing. If the NRC intended to state that all PCR-based genotyping
systems are reliable and generally accepted, surely they would have
simply made that statement instead of painting with such broad strokes.
In any event, as Venegas makes clear in its adoption of the 1992 NRC
Report over the 1996 Report, and as Soto reconfirmed by focusing on the
scientific literature that had developed on RFLP testing after the
publication of the NRC II, the NRC II is not the final word on
California Kelly-Frye DNA issues.
Nor are the issues raised in this case addressed by the recent case of
People v. Allen (1999) 72 Cal. App. 4th 1093. The same misperception of
the science of PCR-based testing evidenced in Wright and Morganti with
respect to DQ alpha and polymarker testing is also evident in Allen with
respect to Short Tandem Repeat testing kits. Without citation to the
Venegas holding that " FBI Agent Lynch, though vigorously defending
the merits of the FBI's RFLP analytical procedures she had followed in
this case, did not purport to be qualified, as a molecular biologist or
otherwise, to testify on questions of general scientific acceptance of
the validity of those prosedures" (18 Cal. 4th at 77), the court in
Allen stated that it failed to see why the testimony of a Cellmark
employee could not alone show general acceptance of "STR's (which
test three [undisclosed] genetic markers.)" Id. at 1097, 1099.
Contrary to all of the cases discussed in section D. 2. , supra, the
court focused not on the specific STR kits employed in the case, but on
the vague concept of "STR testing" in general. In finding that
the People had established through the testimony of a single interested
witness that "STR testing" was generally accepted by an
undefined "scientific community", the court relied heavily on
a Massachusetts case, Commenwealth v. Rosier (1997) 425 Mass. 80 in
which the Massachusetts Supreme Court quoted the NRC II report out of
context for the proposition that
" STR testing is similar in principle to the RFLP (or VNTR) method,
which has been found to be reliable." Id at 1100. In fact, the
portions of NRC II referred to in Rosier and Allen state that STR
methods "are usually capable of unique allelic identification"
(NRC II at 34), that "(a)s more STR's are developed and validated,
this system is coming into wide use" (Id. at 71), and that such
methods " will soon be on the same solid footing as VNTRs"
(NRC II at 117). Such statements are a far cry from the Allen court's
apparent understanding that somehow NRC II had validated all STR testing
platforms for all time. The court cited no California cases standing for
the proposition that one STR based test is the same as any other and
cited no California cases that stated that general acceptance of PCR as
a method to amplify DNA means that all applications of PCR-based testing
is generally accepted as reliable. The court's loose reading of Kelly,
and its failure to even acknowledge Venegas, was apparently prompted by
the court's understanding that "the results generated by a
scientific test once considered valid can be challenged by evidence the
test has since been invalidated", and that "defendant is not
foreclosed from showing new information which may question the
continuing reliability of the test in question or to show a change in
the consensus within the scientific community concerning the scientific
technique." Id. at 1100-1101.
As Venegas clearly holds, an appellate decision upholding one type of
DNA methodology is only binding on another DNA methodology " (i)n
the absence of proof of any material scientific distinction between the
two methodologies." 18 Cal. 4th at 54. Here, the Court has before
it the declaration of Dr. Christie Davis attached hereto as Exhibit D.
Dr. Davis has a Ph.D. in Microbiology and Immunology and is familiar
with the hearing in Allen and with the technology at issue in this case.
Dr. Davis in no uncertain terms lays out the facts that make the tests
before this Court scientifically distinct from those in Allen, thus
allowing this Court to hold a prong one hearing on the Green I and
Profiler Plus kits as well as the ABI Prism 310 capillary
electrophoresis system.
In ¶ 7, Dr. Davis states that the PCR process is not a test itself, but
is a means to increase the number of copies of genetic markers and that
the process itself is complex and varies depending on the genetic marker
of interest. In ¶ 8, Dr. Davis states that the test at issue in Allen,
is different than either of the two STR kits challenged here. It was a
kit produced by a different manufacturer, Promega, and while amplifying
the three of the four markers from Green I, the test used a silver
stained gel rather than capillary electrophoresis. As Dr. Davis notes,
the primers are different than those in the Green I kit, the protocol
used is different than the one for the Green I kit, including the
thermocycler program as well as the number of cycles of amplification.
Id. at ¶ 10.
Perhaps most important is that the results that are obtained from the
Promega kit used in Allen can vary from the results obtained using the
Perkin-Elmer Green I kit. Id. at ¶ 11. Thus one piece of evidence
analyzed with the Green I kit will have a different genetic profile than
one analyzed with the Promega CTT kit. If the Court follows the
reasoning of the People and of Allen, then both results and both kits
would be considered reliable. This is logically inconsistent and would
lead to absurd results. For example, if one lab tested the evidence
using a Promega kit and the results were inculpatory, and then another
lab retested the same evidence with Green I and the results were
exculpatory, then one expert would have to testify that both kits were
reliable and generally accepted. This example illustrates the fallacy of
the argument and the reasoning that STR based tests are interchangeable
because they involve STR markers as a class.
The same reasoning applies to the Profiler Plus kit. As a test, it is
completely different than Green I and the Promega CTT kit. Declaration
of Dr. Davis, at ¶ 12-15. The nine loci tested are different than those
found in Green I and Promega CTT, and include one locus, vWA, that has
produced several papers that "address serious problems that can
ultimately change the genetic profile obtained."Declaration of Dr.
Davis, at ¶ 14. See also, C. Alves, et. al., vWA STR Genotyping:
Inconsistency Between Perkin Elmer's Profiler Plus Kit And Promega's
Geneprint, International Society for Forensic Haemogenitics, Eighteenth
International Congress Abstracts, August 17-21, 1999, San Francisco,
Ca., p. 30 (simultaneous study of vWA locus by the Perkin Elmer's
Profiler Plus Kit and the Promega Geneprint Kit produced an
inconsistency between the genotyping in each kit: using Profiler it was
found to be 18 and with Geneprint 16-18. " Since primer sequences
were not available from the manufacturers we could not sequence the
corresponding regions. However, it is tempting to interpret the
inconsistency as a result of a Perkin Elmer primer annealing
failure...The finding now reported evidences the need for caution when
comparing genotypes or gene frequencies made in amplicons and produced
by different primers."); M.C. Kline, et. al., Nonamplification of a
vWA Allele, J. Forensic Sci. 1998 Jan., 43(1):250(National Institute of
Standards and Technology researcher documents same inconsistency and
indicates that Perkin Elmer " is aware of the problem and they are
actively pursuing an explanation for this allelic dropout by sequencing
the sample"); S. Walsh, Commentary on Kline, MC, Non-Amplification
of a vWA Allele, J. Forensic Sci. 1998 Sept., 43(5) 1103( Perkin Elmer
admits the problem exists, and claims it is caused by a flanking
sequence mutation. "Our laboratory has observed flanking sequence
mutations in several STR loci, including the vWA loci reported here,
D16S539, TPOX. Other laboratories have reported flanking sequence
mutations at D13S317 and DS7820." Perkin Elmer admits that the
problem will continue, but claims that it can be avoided by using Perkin
Elmer products exclusively.) As these articles demonstrate, the primers
in the three kits are different. The reaction conditions are different.
In no way are the three kits interchangeable. The Green I kit could not
be used to test for CTT nor could either the Green I and CTT kits
components be used to test for the Profiler Plus markers. There is no
question that each is an independent and wholly different testing system
that must be evaluated prior to its acceptance.
The logic of the defense position in this case, that acceptance of
Promega CTT does not allow untested acceptance of other STR based kits,
is supported by the People's own witnesses in the Bokin hearing. Dr.
John Tonkyn of the California Department of Justice, testified that
"[t]he kit requires 4.1, developmental validation, before it can be
utilized. In addition, each laboratory that uses this must comply with
Section 4.5, internal validation." (R.T. 5456). Significantly, Dr.
Tonkyn conceded on cross examination that "a validation study which
dealt, for instance, with the Promega CTT system, using the Promega
primers in a silver staining system, would not validate developmentally
or internally the Green I Kit, which uses an entirely different set of
primers, an entirely different software, and entirely different
parameters." (R.T. 5477).
Therefore, the Court has before it through Dr. Davis' declaration and
Dr. Tonkyn's testimony, that general acceptance of one kit does lead to
the general acceptance of another kit. The Green I and Profiler Plus
kits differ in a materially scientific manner from the Promega kit at
issue in Allen. The Court must hold a hearing to inquire into whether or
not the Profiler Plus and Green I kits are generally accepted as
reliable.
Moreover, the method used to visualize the amplified DNA is materially
different than that used in Allen. There, the tests were conducted using
silver stain gels. Declaration of Dr. Davis, at ¶ 16-19. That method is
simple as compared to the capillary gel electrophoresis method.
Capillary electrophoresis is an entirely different technology that is
automated; uses a different polymer system; different chemistry; and
uses two computer programs to analyze the raw data. Id. at ¶16-17. The
systems is known to produce artifacts and no validation studies have
been published. Id.
This Court must take seriously the gatekeeper function delineated in
Kelly and require that this new technology that carries so much weight
with lay jurors be subject to the most stringent examination. This Court
should not be taken in by the illogical and simplistic arguments of the
People that STRs are STRs.
C. THE RULING IN PEOPLE V. MCCLANAHAN IS NOT BINDING ON THIS COURT,
AND THE RULING IN PEOPLE V. BOKIN IS MORE PERSUASIVE AND SHOULD BE
FOLLOWED
The oral ruling by the Honorable Judge Richard Kramer in McClanahan
regarding the Green I kit is not binding on this Court. The extensive
five month hearing in Bokin came after that in McClanahan and is
supported by detailed findings of fact and conclusions of law set forth
in Judge Dondero's sixteen page ruling . (See, Exhibit A). Judge Dondero
ruled that the Green I kit was not generally accepted as reliable within
the relevant scientific community. See, People v. Bokin, SCN 168461,
dated May 5, 1999. As the decision in Bokin makes clear, the standards
not only of the scientific community of molecular biologists and
geneticist, but of the forensic community as delineated in the TWGDAM
and DAB guidelines were not minimally satisfied as regards the
validation and testing of the Green I kit. Therefore, the defendant
respectfully requests that this Court take judicial notice of the
testimony and evidence in Bokin and hold that the Green I kit and the
tests performed in this case using the Green I kit, are inadmissible
under Kelly.
New evidence was presented in Bokin that was not presented in
McClanahan.First, the studies of the California Department of Justice on
the Green I kit were not before Judge Kramer. Those studies indicated
that the Green I kit did not perform as Dr. Cynde Holt testified that
they did. High levels of stutter were observed as well as "reverse
profiles." Additionally, Dr. Holt did not bring to Judge Kramer's
attention the fact that the phenomenon of allelic dropout had been
observed. In particular, Dr. Holt did not inform Judge Kramer that
different results were obtained using Perkin-Elmer's kits, as compared
to the same set of markers manufactured by Promega. This omission was
made despite Dr. Holt's admission in Bokin that she was aware of the
controversy when she testified in McClanahan, and despite the fact that
she stridently urged the court in McClanahan that there were absolutely
no reliability problems with any Perkin Elmer product. Evidence of this
allelic dropout problem was brought out in the Bokin case and is
documented in the articles cited above, one of which (the Walsh article)
is authored by a Perkin Elmer employee.
Additionally, Dr. Holt mislead Judge Kramer in her testimony regarding
the validation studies done by Perkin-Elmer and the forensic community.
In particular, Dr. Holt testified that the California Department of
Justice had validated Green I and that she had personally attended the
talk given by Dr. Tonkyn in which he presented his results. This was
directly contradicted by Dr. Tonkyn and by others at the Department of
Justice. Dr. Lance Gima filed a declaration with the court in Bokin,
stating that the studies conducted by his lab were internal studies and
did not amount to validation of the Green I kit, in part because they
had abandoned the study. Dr. Tonkyn testified that he did not present
any results of his Green I study and that he announced at the talk that
Dr. Holt stated that she attended, that he had not finished the study.
Finally, Dr. Holt testified that certain tests described in the Green I
kit User's Manual, were conducted by California Department of Justice.
In Bokin, Dr. Tonkyn testified that some of the tests described in the
Manual as being conducted by his lab, had not in fact been conducted.
This Court should cast a wary eye upon the order issued by Judge Kramer.
Dr. Holt, the only expert called by the People in McClanahan,
misrepresented evidence and withheld critical information from the
court. Based on her misrepresentation, her credibility is suspect and
should not in a case such as this, prevent this Court from taking a
close and critical look at the Profiler Plus kit, another product of
Perkin-Elmer
III.
DEFENDANT ROBERT NAWI RAISES UNRESOLVED ISSUES PURSUANT TO KELLY-FRYE
As discussed above, several controversies remain that are unresolved
by case law that binds this court. Also discussed above, Venegas clearly
mandates a hearing on the third prong of Kelly to determine whether the
Crime Lab's testing procedures meet scientifically acceptable standards.
Keeping in mind the parameters of case law that binds this court, the
defense submits that the court should hold a hearing to determine
several other issues raised below.
A. THE METHODS USED IN THIS CASE FOR THE COLLECTION, PRESERVATION,
HANDLING, AND PROCESSING OF CRIME SCENE SAMPLES FOR FORENSIC PCR BASED
DNA TESTING --THE DQ ALPHA + POLYMARKER, D1S80, GREEN I, AND
PROFILER-PLUS KIT TECHNIQUES -- ARE NOT GENERALLY ACCEPTED AS RELIABLE
AMONG THE RELEVANT SCIENTIFIC COMMUNITY
1. The Failure of The SFPD Crime Lab to Follow Correct Scientific
Procedures As Outlined In TWGDAM and DAB Guidelines, The NRC Reports,
The Manufacturer's Protocols, And the Lab's Own Protocols Renders Any
DNA Test Results Inadmissible
The prosecution's burden under Venegas is a very heavy one indeed.
There are no California appellate cases approving the methods of the San
Francisco Crime Lab, which uses commercially distributed amplification
and typing kits. The use of a commercially distributed standardized
"kit," however, does not mean a laboratory employs adequate
methods and controls to do reliable testing. Individual laboratory
validation is required, and many scientists, and the NRC itself, believe
that additional scientific controls are necessary to make PCR based
forensic testing reliable. The NRC makes this point clear in its
comments on Roche's DQ-alpha "kit":
One commercial kit for forensic PCR analysis has been marketed. Other
such kits will probably be ready for commercial distribution soon. The
committee sees a potential for introduction of unreliable kits and the
misuse of kits. The existence of a kit suggests ease of use and low
chance of technical error. The committee believes that nonexpert
laboratories will run a significant chance of error using kits. We
therefore recommend that a standing committee (discussed later in this
chapter) consider the issue of regulatory approval of kits for
commercial use in forensic DNA analysis. Even though no precedent exists
for the regulation of tests in forensic DNA applications, we believe
that it might be necessary for a government agency to test and approve
kits for DNA analysis before their actual forensic use.
NRC Report, at 69.
Moreover, the use of PCR based DQ Alpha +Polymarker, D1S80, and STR test
kits, which commenced in San Francisco in March, 1997 in the case of DQ
Alpha + Polymarker and D1S80, in March, 1998 in the case of Green I, and
in February, 1999 in the case of Profiler-PlusFe, is clearly, for the
Crime Lab, at the earliest experimental stages. The forensic use of
these techniques by this lab is barely known to the scientific
community, much less generally accepted.
Furthermore, the prosecution will be unable to demonstrate at the Kelly
hearing that the SFPD has, or employs, a generally accepted method for
the collection, preservation, and handling of crime scene samples for
purposes of PCR-based testing. As indicated above, the NRC II is
emphatic that such controls are necessary becauser "any procedure
that uses PCR methodology is susceptable to error by contamination"
and thus "the integrity of the chain of custody is of paramount
importance." NRC II at 23,25. Moreover, the risk of sample
contamination created by unreliable preservation and handling practices
is seriously compounded by the fact that the SFPD DNA laboratory does
not employ generally accepted methods for identifying, preventing, and
controlling contamination in its processing of samples for PCR testing.
Put simply, the methods employed by SFPD are not generally accepted as
reliable because they create scientifically unacceptable risks of sample
contamination for purposes of PCR testing on the crime samples in this
case.
As recently revealed in Bokin, the failure of the SFPD Crime Lab to
follow correct scientific procedures is substantial.The major
deficiencies can be quickly summarized. The crime lab is not accredited
(see, Penal Code Section 297(c) ["All laboratories, including the
Department of Justice DNA laboratories, contributing DNA profiles for
inclusion in California's DNA databank shall be accredited by the ASCLD/LAB."]),
it lacks critical resources and managerial structure, and until his
recent resignation it is operated by a "technical manager"
(Allen Keel) who was unqualified and operated without adequate
supervision and oversight. The lab has not adopted and enforced sound
and conservative interpretative guidelines and its proficiency and
internal validation testing progam is sporadic and insufficient. The lab
lacks adequate security and contamination control protocols and it
failed to follow industry standards after contamination was found in its
water supply, which required that the lab shut down and run anew all
casework.The lab improperly makes DQA and PM calls in the absence of C
or S dots and fails to use appropriate positive and controls in its
testing.
The Lab also does not have " a managerial staff with the authority
and resources needed to discharge their duties and meet the requirements
of the standards..." DAB Guideline 4.1; TWGDAM Guideline 1.3
According to a letter Mr. Keel himself recently wrote to state auditors,
"the lab has no senior or supervising criminalist overseeing
anything other than Ms. Blake who manages the overall laboratory
operation in a provisional capacity as senior criminalist...The failure
of the audit report to even comment on the disarray of the SFPD
laboratory formal organizational and hierarchical structure is glaring
absent." (Bokin Exhibit 8K).
The evidence in Bokin also convincingly demonstrates that DNA forensic
analysis is a technically demanding and subjective enterprise that
demands specialized training, good judgment, the ability to learn from
one's mistakes, constant continuing education efforts in a rapidly
changing field, impartiality and a proper scientific attitude. Even well
trained PhDs such as Department of Justice employee Dr. John Tonkyn
admitted in Bokin that he had made critical mistakes in interpretation
because of his carelessness and failure to follow sound conservative
guidelines.
The DAB guidelines specifically require that a technical manager such as
Mr. Keel, "shall have at a minimum a Master's degree in biology,
chemistry, or forensic related area and successfully complete a minimum
of 12 semester or equivalent credit hours of a combination of
undergraduate and graduate course work covering the subject areas of
biochemistry, genetics, and molecular biology or other subjects which
provide a basic understanding of the foundation of foresnic DNA analysis
as well as statistics and or population genetics as it applies to
forensic DNA analysis." DAB 5.2.1. The TWGDAM guidelines are
similar but slightly less demanding. See guideline 2.2.3.1
Although Mr. Keel represented to Judge Dondero on direct examination
that he was qualified "pursuant to TWGDAM and DAB guidelines"
(R.T. 2026) it is clear from his cross examination that he does not have
a Master's degree and that he has not yet obtained a waiver of the above
requirement pursuant to DAB guideline 5.2.1.1. The evidence in Bokin
further establishes that Mr. Keel has a B.S. in zoology and that he has
taken an unspecified number of different zoological courses including
immunology, human physiology, and the struction and function of muscle.
He has also taken some early UC extension classes in nucleic acid
biochemistry. That is the extent of his formal training (R.T. 2133,
2135). His CV and his testimony further indicate that his formal
educational efforts stopped arround 1996. His case notes for his
proficiency tests and for the cases before the Court in Bokin indicate
multiple instances of mistakes, shoddy record-keeping procedures, and
failure to follow proper protocol. The cross examination on these issue
revealed many irregularities, including crossed out and changed results,
self grading, and failure to submit the tests on time. The cross
examination also revealed that while employed at the Oakland Crime Lab,
he failed serology proficiency tests because of interpretational and
sample mix-up errors, and committed the same kinds of mistakes in his
DNA casework.
Reviewing all of this evidence, Judge Dondero specifically found that
" Alan Keel's credentials are not in compliance with DAB
guidelines. He lacks the appropriate academic background mandated by DAB
5.2.1 in that he has no Master's degree in the particular sciences
identified in the guideline, nor the semester hours in particular
courses listed. More importantly, he has not complied with the waiver
option pursuant to 5.2.1.1. Familarity with the record involving past
efforts at proficiency testing by Keel suggests lapses in meeting
deadlines. On the record beore the Court, Keel would not satisfy DAB
requirements for any contemporary analysis. This Court cannot be nore
critical of this particular circumstance."Exhibit A at 14
Mr. Keel's overall attitude about the work he does is best summed up by
his response to criticism in a 1995 ASCLD-LAB audit report indicating
that the Crime Lab was not in compliance with TWGDAM Guideline 5.3,
which requires that "a PCR laboratory will require special
laboratory configuration and sample handling" to avoid the ever
present danger of contamination. Keel asserted, contrary to every other
expert who testified in the hearing, that , "THESE SEPARATIONS ARE
RECOMMENDED TO ENSURE PCR SAMPLES WILL NOT BE CONTAMINATED. THESE ARE
RECOMMENDATIONS. THIS WORK CAN BE DONE IN THE SAME ROOM IN A BARN, AS
LONG AS THE -- THE PROPER CONSIDERATIONS AND GUIDELINES ARE IN PLACE TO
CONTROL FOR THINGS LIKE CONTAMINATION. THERE IS NO ABSOLUTE REQUIREMENT
IN REGARD TO DEALING WITH THESE FACTORS." (RT 3275-3276) Despite
the ludicrousness of this statement, Keel stuck with his position upon
further cross examination:
Q. SO YOUR PERSONAL OPINION WOULD BE THAT PCR SETUP, AMPLIFICATION, AND
THE OTHER STEPS IN THE PROCEDURE COULD ALL BE CONDUCTED IN A BARN IN ONE
ROOM WITHOUT ANY PROBLEM WITH CONTAMINATION?
A. I THINK THAT'S POSSIBLE, YES.
Q. AND DO YOU KNOW ANYBODY ELSE IN YOUR PROFESSION, OR IN THE WORLD
THAT SHARES THAT VIEW?
A. I BELIEVE THAT ANYBODY WITH A DEGREE OF COMMON SENSE WOULD SHARE
THAT OPINION.
RT at 3276
Judge Dondero's ruling specifically called this remark "perhaps
the most memorable yet disturbing declaration in the entire
record.". Exhibit A, p. 15 n. 13 As Judge Dondero correctly
concluded, "(a)ny relevant documentation on this topic and
certainly the case law which controls the admissibility of Keel's work
product mandates that DNA analysis is technical and rigorous
analysis....It certainly should not be something so basal that it can be
performed in a barn." Exhibit A at 15.
Keel's carelessness is further reflected in his numerous self-graded
proficiency test record "cross-outs" in which he apparently
changed analytical results after being provided the target values.
Denying such a practice, he admitted that in at least three self-graded
tests there was evidence of cross-outs, and most remarkably he testified
as follows:
MR. BURT: Q. WHAT DO YOU THINK THE CHANCES ARE, MR. KEEL, OF YOU TAKING
THREE PROFICIENCY TESTS AND IN EACH AND EVERY ONE OF THOSE TESTS
COMMITTING TRANSCRIPTIONAL ERRORS, ALL OF WHICH HAPPEN TO CORRESPOND
WITH CHANGING VALUES TO REFLECT THE TARGET VALUES IN TESTS THAT YOU
YOURSELF GRADED?
A. NONE OF THE TARGET VALUES WERE CHANGED, MR. BURT. AND AS FAR AS MY
ERROR RATE IN MAKING TRANSCRIPTIONAL ERRORS, I MAKE TRANSCRIPTIONAL
ERRORS VIRTUALLY EVERY TIME I PICK UP A PEN.
RT at 3969
His past work performance includes an evaluation by his supervisor
that Mr. Keel is "not always as aware of the fundamental
theoretical bases of some of the techniques (he) use(s) as might be
desirable." (R.T. 4603, 4604) The supervisor also observed that Mr.
Keel "occassionally seems to resent technical comments which may
question (his) analytical approach or may contain constructive criticism
from a peer." (R.T. 4606) His attitude toward counsel during Bokin
indicates that this disability is still with Mr. Keel. That attitude, as
reflected in a declaration filed in Bokin, was, as found by Judge
Dondero, " beyond advocacy--it indicated a critical attitude toward
the defense function in a criminal case." Exhibit A at 15 As also
found by Judge Dondero, "(w)hile on the stand, Keel regularly
indicated he chose to ignore particular protocols suggested in the
literature, including that of Perkin-Elmer and peer reviewed
journals." Id at 15 By his own admission, he currently sees no need
to follow suggested remedial advice given to him by federally funded DNA
Consortium auditors who visited his lab in late 1998 who were highly
critical of his practice of failing to sign and date his numerous
"cross-outs."(RT 5087) Clearly, Mr. Keel is not competent to
be the technical manager of the San Francisco DNA Lab or even to be a
competent analyst.
Futhermore, from the moment Bokin began and Mr. Keel submitted an
incredible April 13, 1998 declaration (exhibit 4L) complaining about his
discovery obligations, accusing defense witnesses and counsel of
improper and criminal motives, and misleading the Court about the
results of a December 1997 proficiency test survey, this witness
exhibited a siege mentality that posits that a small "clique"
of misguided scientists and defense counsel are out to get him. As
indicated, Judge Dondero found his bias to be "beyond
advocacy." Exhibit A at 14-15
Mr. Keel's hostile attitude has apparently prompted him to take on the
role of an overzealous advocate. During the course of Bokin, he sat at
the prosecution's table day in and day out, passing notes to the
prosecutors and to witnesses, and he did this despite his complaints to
Judge Dondero and to the media that he is overwhelmed with case work.
His hostile attitude while testifying was evident throughout the hearing
and his hostility extends even to his own supervisor, Ms. Blake, who was
forced to apologize for his conduct to state auditors in a letter in
which she wrote that she, "must apologize a bit for his tone. He is
totally overwhelmed with work and angry about our circumstances."
(exhibit 8K) The comment by Keel that prompted this apology was his
assertion that, "the lab has no senior or supervising criminalist
overseeing anything other than Ms. Blake who manages the overall
laboratory operation in a provisional capacity as senior criminalist...the
failure of the audit report to even comment on the disarray of the SFPD
laboratory formal organizational and heirearchical structure is glaring
absent." Id.
The bias he exhibits in his casework is of a different and more
dangerous sort. The evidence before the Court in Bokin indicated that
Mr. Keel has not established or adhered to the "sound conservative
guidelines" touted by Dr. Holt, but rather approaches each case on
an ad hoc basis and manipulates the analytical procedures until a
desired result is obtained. His analysis in the Franklin and Long case
in which he reran tests over and over again until he got the desired
result is the best example of this bias. (R.T. 3210-3215) As indicated
in NRC II, "bias in forensic science usually leads to sins of
ommission rather than commission possibly exculpating evidence might be
ignored or rejected. Contradictory test results or evidence of sample
mixture may be discounted. Such bias is relatively easy to detect if
test results are reviewed critically." NRC II pg. 84-85. Test
results were reviewed critically in Bokin and the hearing in that case
demonstrates that Mr. Keel's bias renders him incapable of fairly
conducting DNA analysis.
Moreover, although Judge Dondero ultimately ruled that " the
prosecution only barely satisfied prong three" (Exhibit A at 16)in
the cases before the Court, he also concluded that his critical comments
on the lab should " serve as a forewarning", and that
"the prosecution should see that problems identified in Tab 32, as
well as the Audit Reports and Grand Jury study are finally
resolved." Exhibit A at 15-16 Tab 32, an annotated summary of
Keel's testimony showing his numerous departures from the TWGDAM
Guidelines, DAB Standards, the manufacturer's protocols, and Keel's own
protocols is attached hereto as Exhibit E. The Audit reports before the
Court were the 1995 ASCLD-Lab audit and a 1998 audit report by the State
Board of Auditors, which Mr. Keel himself described as superficial. The
Grand Jury study before the court was a 1995 civil grand jury report
which was highly critical of the Lab.
Significantly, however, what was not before the Court were the results
of a December, 1998 audit conducted by the federally-funded DNA
Consortium, which was designed to be a full audit to assess compliance
with TWGDAM Guidelines and DAB Standards.This audit report, which was
not issued until April 7,1999 and did not become available to the
defense until May 25,1999 was not available in time for Judge Dondero's
May 6, 1999 ruling. The report, and the lab's response to it, are
attached as exhibit B. The report removes any doubt as to whether the
Crime Lab can satisfy prong three in Mr. Nawi's case. Although the
report must be read in its entirety to appreciate all of the
deficiencies identified by the auditors, it is important to note that
the auditors found numerous deficiencies, including the following
substantive problem areas:
1.Insufficient space
2.Inadequate number of personnel for caseload
3. A policy allowing casework with materials not previously QC-checked
4. A DQA1/PM interpretation policy unsupported by external validation
5. Lack of adjudicated-type samples for internal validation of Green STR
I
6.Initiation of casework prior to authorization of methods
7. Inadequate documentation of technical and administrative reviews,and
8. Inadequate audit trail for equipment and materials used in testing
In regard to item 6, initiation of casework prior to authorization of
methods, the auditors specifically rejected the rationale put forth by
Keel in Bokin to the effect that he did not have to undergo any internal
validation in San Francisco because he had already done so in Oakland.
The auditors explained that this reasoning was "problematic":
The purpose of internal validation(also known as in-house) validation is
to demonstrate acceptable performance of the total system: analysts,
facilities,equipment, and materials. PCR systems can be sensitive to
differances between particular instruments, water quality, and ambient
temperature, among other things. One member of the audit team (John
Hartman, also a member of the TWGDAM committee that drafted the
Guidelines) states that such a transfer of validation was not invisioned
by TWGDAM, hence the adjectives internal and in-house. In defense of
SFPD's position, it should be pointed out that even if its internal
validation had taken place within its doors and using its own
equipment,over time all of the personnel, equipment, materials and even
facilities could be changed without its formally needing to revalidate
following each change.
Exhibit D at 8
In any event, the auditors specifically concluded that "(t)hese
deficiencies and those of the documentation system should be addressed
before the program implements Profiler-Plus and CoFiler testing. A
follow-up external audit at that time is highly recommended." Id at
16. To the defense's knowledge, after the auditor's visit, Mr. Keel
resigned and no such follow-up audit has been conducted. Moreover, the
Lab's response to the audit indicates, per the pattern established in
Bokin, that Mr. Keel refuses to even acknowledge that any defiencies
exist, much less that he plans to take any action to remedy them. On
this record, the "bare" preponderance found present in Bokin
is totally lacking here as to prong three of the Kelly test.
2. The Reason That Forensic PCR Testing Has Not Achieved General
Acceptance In The Scientific Community Of Molecular Biologists Is
Because Serious Problems Have Been Encountered. There Are At Least Six
Flaws In The Application Of Current Forensic PCR Testing Methods:
Contamination, Degradation, Misincorporation, Differences In Qualitative
And Quantitative Fidelity, Differential Amplification, and False
Positives.
The PCR technique was first developed by Dr. Kary Mullis while
working for the Cetus Corporation in 1985. Dr. Mullis won a Nobel prize
for this work.See generally, Kary Mullis, Dancing Naked in the Mind
Field (1998); Paul Rabinow, Making PCR: A Story of Biotechnology (1996).
Cetus owned a patent on the PCR technique which survived a lawsuit by
DuPont, which claimed the technique was not patentable. Then, Cetus, and
the PCR patent, were purchased by Hoffman-LaRoche (Roche). As indicated
in n. 15 supra, on December 7, 1999, in Hoffman-La Roche, Inc. and Roche
Molecular Systems, Inc v. Promega Corp __F. Supp. __ (N.D. Cal. Dec. 7,
1999), a patent infringement case, Federal District Court Judge Vaughn
Walker recently issued a sweeping ruling in which he concluded, based on
extensive trial testimony, that Roche's patent on the AmpliTaq® DNA
polymerase enzyme must be invalidated because of Cetus's numerous
instances of fraudulent conduct in securing the patent. Significantly
for this case, the instances of misrepresention involved Cetus and Roche
scientists lying about in- house scientific studies.
The PCR technique is used widely in research laboratories and many
clinical laboratories, most notably immunological laboratories doing
clinical work on organ transplants. Nonetheless, the scientific
community has consistently expressed caution about the transfer of PCR
technology to forensic applications. For example, the Office of
Technology Assessment of the United States Congress reported in July,
1990:
At present, however, the enthusiasm of some for PCR applied to forensic
casework is tempered. Cautionary voices warn that, compared to RFLP
analysis, all the possible artifacts and steps necessary to avoid them
have not been fully identified. Some believe that additional studies of
PCR on simulated or real samples is necessary to ensure that problems
often encountered with real samples, including DNA and non-DNA
contaminants, do not interfere with accurate PCR use in forensic
applications.
More recently, the NRC I began its discussion of forensic PCR testing
with this observation:
PCR is a relatively new technique in molecular biology, having come into
common use in research laboratories only in the last four years.
Although the basic exponential amplification procedure is well
understood, many technical details are not, including why some primer
pairs amplify much better than others, why some loci cause
systematically unfaithful amplification, and why some assays are much
more sensitive to variations in conditions. Nonetheless, it is an
extremely powerful technique that holds great promise for forensic
applications because of its great sensitivity and the potential of its
use on degraded DNA.
And the NRC ended its discussion of PCR by concluding:
PCR analysis is extremely powerful in medical technology, but it has not
yet achieved full acceptance in the forensic setting. The theory of PCR
analysis, even though it is the analysis of synthetic DNA as opposed to
the natural sample, is scientifically accepted and has been accepted by
a number of courts. However, most forensic laboratories have invested
their energy in the development of RFLP technology and have left the
development of forensic PCR technology to a few other laboratories.
Thus, there is no broad base of experience in the use of the technique
in identity testing. [emphasis added]
The reason that forensic PCR testing has not achieved general
acceptance in the scientific community of molecular biologists is
because serious problems have been encountered. There are six basic
difficulties with the application of current forensic PCR testing
methods: contamination, degradation, misincorporation, differences in
qualitative and quantitative fidelity, differential amplification, and
false positives.
a. Contamination
Contamination is the single greatest problem in the transfer of PCR
technology to forensic testing. This is because the extraordinary
ability of PCR to reproduce a single copy of DNA is also its greatest
disadvantage. "PCR is not discriminating as to the source of the
DNA it amplifies, and it can be exceedingly sensitive." (NRC I at
65). " Any procedure that uses PCR is susceptable to error caused
by contamination leading to amplification of the wrong DNA." (NRC
II at 71)
Contamination can arise in a myriad of ways. The general categories of
contamination outlined by the two NRC reports are:
a) Contamination from handling in the field during collection, either by
cross-contaminating samples with each other directly,
cross-contaminating samples through DNA carry-over on evidence gathering
instruments or the gatherer's clothing, mixing up samples, or
inadvertent contribution from the biological products from the evidence
gatherer him or herself (sweat, sneezing, dandruff, etc.). "The
important consequences of those sorts of contamination are that samples
might appear to be mixtures from several persons and, in the worst case,
that only the contaminating type might be detected. The concern is
greater with PCR-based typing methods than with VNTR analysis because
PCR can amplify very small amounts of DNA. A false match could occur if
the genetic type of the contaminating materials by chance matched the
genetic type of a principle (such as a suspect) in a case or worse, if
the contaminant itself came from a suspect in the case." (NRC II at
83). In this case, Allen Keel testified in Bokin that he does not get
involved in monitoring the procedures for evidence collection because
such procedures are implemented by crime scene investigators not subject
to his control( See Exhibit E, summary of Keel testimony at 16-17);
b) In the laboratory, cross-contamination of samples to each other, or
contamination emanating from the evidence handlers and their
instruments, while samples are being manipulated, sorted, labeled,
dried, cut, before, during, and after the DNA extraction process.
"For example, in the loading of an electrophoresis gel, a sample
loaded in one lane might leak into an adjacent lane, which might then
appear to contain a mixed sample. Confusion resulting from lane-leakage
problems is typically avoided by leaving alternate lanes empty or by
placing critical samples in nonadajacent lanes, and this should always
be done." (NRC II at 82) In this case, the evidence in Bokin
established that Mr. Keel has had numerous lane-leakage and broken test
tube problems and that "as a matter of convenience" he does
not leave alternative lanes empty. (Exhibit E at 14, 24) The evidence
also showed that he does not follow the recommendation of the NRC II to
seperately test evidence and reference samples.(Id. at 8) It also
established that Keel does not believe in, and therefore does not
follow, the recomendations of the F.B.I. and others that DNA lab
technicians should wear protective hair nets and face masks.(Id. at 6).
It further established that he uses scissors to cut evidence and then
cleans the scissors with ethanol or water, a procedure all other experts
opinioned was totally inadequate to control for examination. (Id. At
5-6) As indicated above, Keel incredibly believes and acts on the belief
that PCR contamination is not a concern and that PCR testing can be
carried out reliably in a barn. (RT 2378);
c) Contamination from contaminants in solutions, reagents, aerosols,
water, and even in Taq DNA polymerase. As the NRC I warns, "(e)ven
the simple act of flipping the top of a plastic tube might aerosolize
enough DNA to pose a problem." (NRC I at 66). The NRC II recommends
that " adherence to a standard QC program provides safeguards
against these kinds of laboratory error." (NRC II at 82). In this
case, two of the specific criticisims of the recent December 1998
federally sponsored audit were that the SFPD DNA lab had " (a)
policy of allowing casework with materials not previously
QC-checked", and that the lab had "(i)nadequate audit trail
for instrumentation and material used in testing."( Exhibit B at p.
2). The evidence in Bokin also established that the lab routinely
ignored indications of contamination in its positive and negative
controls, that it keeps more than one test tube open at a time contrary
to the kit manufacturer's protocol, and that at one point, prior to the
testing in this case, the lab discovered contamination in its water
supply and yet, contrary to standard practice, the lab continued to
operate instead of shutting down and redoing all casework conducted to
date. (Exhibit E at 6-9, 17-19).One article produced at the hearing
(See, section III A. 2.c.iii, # 5, infra)indicated that contamination
has been documented even in Taq DNA polymerase, which is the essential
ingredient in all the labs PCR reactions. Mr. Keel testified he was
unaware of the article or the problem. (RT 3096);
d) PCR product carryover contamination - the contamination of evidence
samples or reaction solutions with PCR products from prior
amplifications. "The most serious problem is contamination of
evidence samples and reaction solutions with PCR products from prior
amplifications." (NRC I at 66). " A false match can occur if
the genetic type of a contaminant matches by chance the genetic type of
a principle in the case; in the worst case, the contaminant originates
from another party in the case." (NRC II at 84). The NRC II
recommends among other precautions the maintenance of a one-way flow of
material and the use of chemical sterilization precautions. In this
case, the SFPD follows neither of these recommendations and the federal
auditors specifically found that although " post-PCR manipulations
are restricted to a single room...photographs of PCR product gels and
dot-blot strips are returned to the DNA extraction room for storage in
three-ring binders with other analytical records." (Exhibit B at
13); and
e) Mixed samples -- contamination from the fact that a crime scene
sample, such as blood stain, could be a mixture of bloods or other
biological fluids or substances. NRC I at 65-67; NRC II at 84. The NRC I
states that "(a)s a rule, mixed samples should be interpreted with
great caution....Interpretations based on quantity can be particularly
problematic--eg., if one saw two allelles of strong intensity and two of
weak intensity, it would be improper to assign the first pair to one
contributor and the second pair to a second contributor unless it had
been firmly established that the system was quantiatively faithful under
the conditions set." (NRC I at 66). In this case, as is set forth
more fully below, Keel can include mr. nawi as a possible contributor to
a twelve year old fingernail sampling only by "inferring" the
DQ alpha type of the sample based on a dot blot intensity analysis that
is not even endorsed by Keel's then reviewer and now employer Dr. Blake.
In sum, no matter where one turns in this case, there are major
methodological flaws in the PCR testing performed by Allen Keel and his
assistants.
Recognized leaders in the scientific community, unaffiliated with
laboratories who have a vested interest in the immediate implementation
of current forensic PCR methods, have urged against the use of such
testing until more stringent controls can be employed to guard against
sample contamination, including independent blind proficiency testing of
laboratories. The views of these scientists alone preclude a finding of
general acceptance.
First, consider the observations of Dr. Richard Lewontin, a pre-eminent
geneticist who has been profoundly influential in the scientific debate
over forensic DNA testing:
The usual scenario in a criminal trial is that a small amount of
dried blood, semen or tissue scraping is recovered from the scene of a
crime. A suspect is then identified, and a sample of his or her blood is
taken on cubic-centimeter amounts. DNA from the crime scene scraping and
from the suspect's blood sample are then compared side by side in the
same laboratory. (Sometimes the tissue or dried blood is found on
property of the suspect and it is the victim's DNA that is to be matched
with it.) While there is more than enough DNA recoverable from the
suspect's large blood sample to carry out the needed procedures, the
very small, and often degraded, sample from the crime scene does not
contain sufficient DNA for the comparison. To obtain sufficient
material, the DNA from the crime scene is "amplified," that
is, copied thousands or millions of times in a procedure known as
polymerase chain reaction (PCR). As the name of the procedure suggests,
the original small number of DNA molecules are copies once, then these
copies plus the original are copied a second time and so on for a number
of cycles, increasing the total population of molecules exponentially
until a sufficient amount has been produced for the matching procedure.
The problem with the PCR technique is that because of its chain nature,
contaminant molecules in the original sample may also be amplified and,
since the original crime scene sample contained so few molecules,
contaminants may overwhelm the original in the amplification. In
addition, small differences in DNA sequence can have very large effects
on the relative amplification of the components in an original mixture.
Now consider the actual practice in a forensic DNA laboratory. A
technician is handling two samples. One is the very large DNA sample
from the suspect's blood, the other is the minuscule DNA sample from the
crime scene, which is then amplified by PCR. The situation is ideal for
PCR contamination, with the result that the suspect's DNA will not
really be compared with that from the crime scene, but with his or her
own DNA that has just been replicated in the PCR reaction. The result
will be a perfect match.
All of us who use the PCR technique regularly are acutely conscious of
the contamination problem, and the best laboratories have suffered
occasionally from it. The perspiration and "oils" on
fingertips have provided enough DNA contamination in PCR experiments to
give completely artifactual results. Only careful replication catches
these errors, and some errors have not been caught until much later when
another laboratory found conflicting results. In the forensic context,
where the liberty and even life of the suspect is in question, it is
essential that courts be assured that laboratories are taking careful
precautions against these contamination errors, not to speak of grosser
errors of recording, etc. Representatives of commercial laboratories
that have previously been found to provide erroneous results have told
interviewers that they have "cleaned up their act." Perhaps
they have, but we cannot know without independent checks, and anyway
what about the people convicted before they "cleaned up their
act"? The FBI has consistently refused to allow independent quality
control checks, relying on their own internal procedures.
In a forensic context, where the liberty and even life of a suspect is
at stake, there must be frequent, independent and unannounced
inspections and tests of DNA laboratories, on the model of the
inspections carried out by radiation safety officers and the Department
of Energy of laboratories using radioactive materials. The issue is
certainly too important to be dismissed by the unsubstantiated opinion
of someone not familiar with the technical procedures at first hand. If
the data themselves are unreliable, questions of probabilities of
alternative suspects are irrelevant.
R. C. Lewontin, Comment: The Use of DNA Profiles in Forensic
Contexts, in "DNA Fingerprinting: A Review of the
Controversy," 9 Statistical Science, 259-260 (1994).
Dr. Richard J. Roberts won the Nobel prize in genetics for his work in
the area of restriction enzymes. He has testified numerous times as a
prosecution expert in cases involving RFLP analysis and traditional
Southern blotting methods. Dr. Roberts has submitted a number of
affidavits opposing the use of forensic PCR analysis. In an affidavit
submitted for consideration at a Frye hearing in Washington v. Gentry,
he had this to say:
First, I will make some general comments about PCR technology. The
technique is an extremely powerful one. In principle, and occasionally
in practice, a small region of a single DNA molecule can be amplified
sufficiently to allow its detection. This means that DNA from a single
cell in principle can be detected by the method. It is this exquisite
sensitivity that makes it desirable for use in a forensic setting where
often only limited amounts of material may be available. Unfortunately,
it is this very sensitivity that makes it also prone to the possibility
of contamination. This can result in the amplification, not of the small
amount of sample that was intended, but rather of a small amount of an
unintentional contaminant. Indeed, problems of contamination are a major
source of error in the laboratory and extraordinary precautions need to
be taken when performing multiple PCR experiments to avoid amplifying
samples that have become contaminated through aerosols or sloppy
laboratory procedures. Of course, knowing that such contamination can
take place means that when working in a laboratory one takes the
necessary precautions to avoid it. Unfortunately, the samples that are
collected from crime scenes are, by their very nature, usually
contaminated to begin with! It thus becomes extremely important that
proper checks and controls are carried out to be certain that the
product amplified from a particular sample is from the sample and not
from the contaminant. How can this be done? Unfortunately, to the best
of my knowledge there is no generally accepted method by which this can
be guaranteed at the present time. The reasons are described below.
PCR is a technique that is still relatively new. It has only been in
routine use in academic laboratories during the last two years and so
our experience with the method is itself fairly limited. This experience
for the most part has involved the pristine academic laboratory setting.
Much less experience is available for forensic samples. In a laboratory
setting it is usually possible to take the great precautions that are
necessary to avoid dealing with mixed samples. Nevertheless, it occurs
and occasionally amplification results in bands other than the ones
expected. In the laboratory one can usually go back and repeat the
experiment, perhaps changing reagents or merely being more careful about
sample handling. That such artifactual amplification will happen from
time to time is to be expected.
Bokin Exhibit 27,Roberts Aff., February 22, 1991, Washington v.
Gentry, at 2-3 [emphasis added].
Finally, and perhaps most significantly, there is the position of Dr.
Kary Mullis himself, the man who developed the PCR technique and won a
Nobel prize for doing so.
Dr. Mullis believes that the PCR DQ-alpha test should be used for
exclusions but not inclusions. He has publicly taken this position, and
criticized the inadequacy of controls being used to protect against
contamination, in two Kelly hearings in the past two years. See, Bokin
Exhibit 28, Mullis testimony, People v. McIntosh & Schlaepfer and
People v. Moffet.
In the course of his testimony in People v. McIntosh & Schlaepfer,
Case No. 33026, Tulare County Superior Court, on August 23, 1994, Dr.
Mullis touched upon the major issues the defense is raising here.
First, Dr. Mullis specifically addresses why special care must be taken
in the collection, handling, and preservation of crime scene samples
when PCR testing is going to be done, and why such procedures must be
considered a part of the forensic PCR testing method:
Q. Okay. What I'm pointing at, for the record, is the DNA technology in
Forensic Science by the National Research Council that was published in
'92. I forgot what exhibit it is.
But, in any event, you are familiar with some of the findings and the
information that's contained in that report they submitted in '92?
A. Yes.
Q. Looking at page 52 where they distinguish between the DNA diagnostics
and DNA forensic where they give descriptions, I'm just curious whether
this accurately represents what you're describing here when you're
talking about diagnostics. You're usually involving clean tissue
samples, known sources, repeated -- usually can be repeated comparison
of discrete alternatives. Is this the kind of stuff that you're
acquainted with?
A. Right. You have control over everything. Not over everything, but
over many more things when you're in a diagnostic situation than if
you're just selecting samples that have been laying around.
Q. So where they describe the forensic differences that are involved in
forensic typing, they're speaking of degraded and contaminated
possibilities, multiple unknown sources, inability to repeat always,
matching of samples that can be from a wide range of alternatives. These
are the kinds of things that you're aware of?
A. Yes, it's much more difficult to know that your results are really
what they seem to be in that situation. I think -- criminalists, I
think, are the people that go around and collect that stuff. They
probably will get better at it. They'll eventually know what you can and
can't do. But it's a vast problem compared to what you have in a clinic
or in a hospital.
McIntosh, Mullis at 2340, line 17 to 2342 line 3.
Q. I didn't ask the question very well, I'm sure. But I think you
answered it. Just to make sure though, what I'm talking about is that
let's say you have samples that are picked up at a crime scene and not
handled carefully and contamination could occur, a cross-mixing of
various samples of DNA. Because of PCR's inherent sensitivity, if you
ran PCR procedure on that DNA and then attempted to type it using the
dot blot method, for example, would it be-- tend to be more sensitive to
that error because of the evidence that was handled in PCR versus some
other type of typing?
A. Yeah, I think I answered saying in certain cases when you wouldn't
always know when that was, it certainly would be. In other cases, it
introduces an uncertainty there that you don't have it you don't
amplify. Because if you don't amplify and you get a thousandth part
contamination, you know it will still be a thousandth part when you get
to the analysis. It will be a wispy little thing somewhere and you'll
say that's probably an artifact.
If you're going to amplify it and you pick up a contaminant that happens
to be a lot easier to amplify than the sample that you really want to
amplify, then that could become a real problem, because ten cycles gives
you a thousand fold amplification, 20 cycles gives you a million, 30
cycles, which these people are doing, 30 or 32, gives you a billion.
Q. Thank you. Now, you said something this morning when you were talking
about the application of PCR to the research and diagnostics area and
then you were asked about the application in the forensics area. I think
you mentioned that forensics is not the same as research and
diagnostics. And then you said the certainty level is a lot lower, and
I'm not sure what you meant by that.
In other words, when you're using PCR in the forensic area, is the
certainty level lower than you would expect it to be?
A. I don't think the certainty level of the PCR reaction inherently
would be any lower. It doesn't matter what it's being used for. The
certainty level of the whole analysis involves the collection, the
maintaining, the processing of the samples.
McIntosh, Mullis at 2387 line 2 - 2388 line 24.
Similarly, Dr. Mullis decisively rebuts a common prosecution contention
that PCR-based tests must be assumed to be free of contamination if the
negative amplification control (the "blank control") that
comes with the Roche kit does not light up. Echoing the NRC's warning
that a blank control will not necessarily reveal contamination ("[E]ven
in a laboratory contaminated with PCR carryover, blank controls do not
necessarily become contaminated on every occasion." NRC 1 at 67),
Dr. Mullis describes the dimensions of the problem:
Q. Now, you said earlier something, I won't go into the details, but it
is listed here on page 67, that blank controls do not necessarily become
contaminated on every occasion. This is a problem?
A. That's a tough problem, too. It is. How many blank controls are you
going-- like I said, you cannot ever show that there wasn't one
contaminating molecule in a bottle without taking the whole bottle and
putting it into the PCR reaction. Once you've done that, you got to make
another bottle. Just like you can't ever know if in fact there's just
one, you know, virus in your blood stream if you're looking for a
dangerous virus. You can't sample the whole thing. You don't sample the
whole bottle of a particular reagent when you take a sample of it.
If it's highly contaminated, the statistics say you stick your pipette
in there and take it out and you would get at least one of the
contaminants. If there's only ten or fifteen in there, it will only be
ten or fifteen times when you're using that bottle that you end up
pulling one of them out. You never know when that would be. So you have
to have five controls and they come out blank and then the unlucky time
you might stick it down in there and put it into one of your
experimental tubes and it will show up there, but not in any of your
controls.
McIntosh, Mullis at 2375 line 21 to 2376 line 23.
On cross examination Dr. Mullis remained firm in his contention that
forensic PCR tests were not as reliable as other diagnostic clinical
applications, particularly when analysis was being performed on degraded
samples.
Q. As far as the actual basic science of the diagnostic laboratories
using PCR, that is no different than a forensic lab using PCR, correct?
A. The basic science is basic science. I suppose the problems in the
forensic lab, we have tried to list some of the ones that are in a
forensic lab where it's very difficult or impossible to even make a good
shot at solving them. Where in a clinical setting it is possible to take
precautions that make your results much more reliable. That's what I was
talking about just before the break which had to do with being able to
know exactly how much DNA you're adding, how many copies of what it is
you're looking for to expect and to set up the conditions of the
amplification in accord with that.
If you have a sample that's degraded, you don't know how much DNA's in
it, you just have to take a shot at it. And you might go too many cycles
and, therefore, bring a contaminant up to a level that it will look like
if it wasn't a contaminant.
McIntosh, Mullis at 2420 line 1 to line 24.
Finally, Dr. Mullis rejected the idea that combining, as the prosecution
has done in this case, several different DNA systems would make the PCR
testing any more reliable:
Q. In a case where we use several different systems, one including DQ-alpha
as a site, we use RFLP markers as other sites, conventional serological
evidence as other sites, would the combination of all those sites be
useful?
A. If there weren't so much trouble with reliability on the DQ-alpha
thing, that would certainly be data that you could add up together and
say well, the probability based on DQ-alpha is that he's in a 17 percent
group. The probability based on blood typing is that he's on a 30
percent group. The probability that you do with some RFLP or a couple of
those puts him in an even smaller group. But you still would come up
with maybe a number of, say, two out of a hundred people might or two
out of a thousand or one out of a million maybe, if you're lucky.
Depends on the data that you collect, whether it's rare data or whether
he's got common things at those places.
But the reliability of the RFLP data, it says why add that to this whole
mishmash of stuff that you've got? Why take some piece of evidence that
you can find plenty of scientists will come in and say I wouldn't trust
it on myself? I would repeat it over and over. I've got this problem
with contamination. I've got this one with differential amplification.
This method relies on too many things that are easy to get wrong to make
it really worth pursuing. That's all I'm saying. I'm saying why pick a
method, why use a method that is so subject to human error and to just
things that are way beyond our control? Why pick a method like that to
use in a Court of Law? You wouldn't use that on your taxes.
McIntosh, Mullis at 2450 line 12 to 2451 line 21.
Other molecular biologists from the academic community and the clinical
community have expressed similar objections to the concerns voiced by
Drs. Lewontin, Roberts, and Mullis. For instance, Dr. Ashok Bhagwat, a
professor in the Chemistry Department at Wayne State University, had
extensive experience studying molecular biology at Cold Spring Harbor
where he received his post doctoral training working under Dr. Richard
Roberts. In Washington v. Gentry, Dr. Bhagwat offered these succinct
views:
Q: Dr. Bhagwat, is PCR technology generally accepted in the scientific
community for use on crime scene evidence?
A: No, it's not.
Q: And could you describe to the Court some of the general reasons why
you have that opinion?
A: There are several problems with the material that you are likely to
obtain at a crime scene. It is most likely to be contaminated. And let
me just slightly digress and say, PCR is so powerful that a single cell
contamination could be devastating to the result, and hence,
contamination is a big potential problem and it is unlikely that it is
easily solvable for evidence arising from crime scenes.
Bokin Exhibit 29, Washington v Gentry testimony, 28 RP 2238-2239.
Dr. John Gerdes is clinical director at Immunological Associates of
Denver, a reference laboratory that uses the same PCR technology
involved in this case in conjunction with transplants (matching organ
donors and recipients) and screening for AIDS and other infectious
diseases. He has testified over twelve times expressing, from the
hands-on perspective of one who runs a clinical laboratory, the same
objections raised by Drs. Lewontin, Roberts and Mullis. Dr. Gerdes
testified most recently in the McIntosh case, in tandem with Dr. Mullis,
and the defense submits this transcript for the Court's edification.(See,
Bokin Exhibit 30).
Most of the discussion so far has focused on cross-contamination of
forensic samples in their collection, preservation, handling, and
processing. Before leaving the subject of contamination, however,
special mention must be made of the phenomenon of PCR product
contamination: contamination of samples by PCR product from previously
run reactions in a laboratory. This is sometimes called "carryover
contamination." One of the developers of the Cetus/Roche kit, Dr.
Russell Higuchi, describes the phenomenon:
More unique to PCR is the possibility of carryover contamination from a
completed PCR to another sample yet to be amplified. Because by the
nature of PCR, PCR product will seed production of more PCR product, the
sheer number of copies of PCR product after amplification can make the
consequences of such contamination more dramatic. A typical PCR could
have 1012 copies of an amplified gene. If a preparer inadvertently
transfers, as before, 0.1 l of PCR sample A into sample B, even though
sample B has a relatively high concentration of human DNA, the number of
copies of the target, single-locus gene that derive from sample A far
outnumber the copies that actually stem from sample B. Thus, the DNA
type obtained will be that of A and not B, and the relative amount of
the B type is so small that it would not even show up in the test,
eliminating the possibility that the presence of more than two alleles
would flag the occurrence of the contamination.
The NRC highlighted the problem of PCR carryover contamination as a
key area of vulnerability, declaring "it has become clear that
carryover products from one PCR reaction to another must also be
eliminated." NRC 1 at 67. Accordingly, the NRC commented that
"[m]ethods of detecting and preventing contamination from one PCR
reaction to another in forensic laboratories are generally still in
their early stages, and additional development should be
encouraged."
It is plain, however, that forensic PCR laboratories have not responded
appropriately. For example, the NRC pointedly suggested that it would be
a good idea to treat all evidence samples with uracil N-glycose (UNG)
before amplification to destroy any PCR carryover from previous PCR
reactions. Id. Clinical laboratories have followed up on this idea. Dr.
Gerdes reports that his and other clinical laboratories "don't do
anything without UNG anymore," McIntosh, Gerdes, 1915 line 6, and
that the UNG control was featured in a PCR chlamydia kit which won
recent FDA approval. Id., 1914 line 19. Forensic PCR laboratories,
including the one which did testing in this case, do not use UNG or any
other chemical treatment to protect against carryover contamination.(RT
3012-3013)
Most importantly, since PCR carryover contamination can be a systemic,
cumulative problem in a laboratory, the NRC demanded that controls must
be in place to monitor "general contamination" in forensic PCR
laboratories:
In view of the problem of contamination due to handling and carryover,
laboratories must incorporate contamination control into their standard
operating procedures. And outbreaks of contamination and the steps taken
to correct the problem should be documented.
NRC 1 at 67.
In this case, as indicated above, in December 1998, two federally
sponsored auditors concluded after careful inspection that the SFPD DNA
lab had " (a) policy of allowing casework with materials not
previously QC-checked", and that the lab had "(i)nadequate
audit trail for instrumentation and material used in testing." (
Exhibit B at p. 2).The evidence in Bokin also established that the lab
routinely ignored indications of contamination in its positive and
negative controls, that it keeps more than one test tube open at a time
contrary to the kit manufacturer's protocol, and that at one point,
prior to the testing in this case, the lab discovered contamination in
its water supply and yet, contrary to standard practice, the lab
continued to operate instead of shutting down and redoing all casework
conducted to date. (Exhibit E at 6-9, 17-19).
b. Degradation
One of the reasons both TWGDAM and the DAB Standards require extensive
developmental and internal validation of each PCR testing platform is to
ensure that each platform can reliably type specific kinds of samples
under a variety of environmental circumstances encountered in crime
scene samples. TWGDAM Section 4.1.5.6 specifically requires that the
required developmental validation must include environmental studies
"so that the effects of factors such as matrix, age, and
degradative environment (temperature, humidity,UV) on a sample are
considered." See also, DAB Standard 8.1.2.2 (The developmental
validation shall include... species specificity, sensitivity, stability,
and mimixture studies...").
In this case, the Crime Lab purports to be able to accurately type
thirteen year old fingernail and victim reference samples allegedly
containing DNA profiles of at least three persons. As indicated above,
before such testing is admitted into evidence, it is generally accepted
that developmental and internal validation must be conducted to
demonstrate that typing of aged fingernails is possible. Although the
scientific community has begun to examine this issue, the existing data
is fragmentary and raises more questions than it answers. It is far to
early to conclude that typing of aged mixed fingernail scrapings is
generally accepted as reliable.
Even as to the typing of fresh fingernails, problems have been
encountered because of the unique nature of fingernails. So far, it has
been possible on fresh samples to obtain only highly degraded DNA. One
recent article discussed below proposes the following possible
explanation:
Although DNA profiles are currently reproduced from different
biological sources( hair roots, semen, faeces), keratin cutaneous
appendages have been traditionally regarded as refractory to DNA
extraction and genomic analysis....Nails, as other cutaneous structures,
derive from hyperkeratinization of epidermal cells... A tentative
explanation for the highly degraded status of harvest DNA...could be in
the apoptosis ( cellular programmed death, associated to nuclear DNA
fragmentation) to which keritinized tissues undergo. To follow their
psysiological fate, nail cells are likely to activitate specific
biochemical pathways, leading to nuclear membrane disintegation and
keritin content increase. In this process, the DNA is set free in the
cytoplasm, it is attacked and psysiologically degraded by nucleases.
C. Capelli, et. al., Fingerprints From Fingernails?, Advances in
Forensic Haemogenetics 7 (1997) p. 100-102.
To date, other studies considering the problem of degradation in
general or the typing of fingernails specifically have been equally
tentative, or, if not, have indicated definite problems in typing
degraded material. A fair overview of this literature is as follows:
1. L. Potsch et al, HLA DQ alpha Typing of Human Fingernails,Advances
in Forensic Haemogenetics 4 (1991) 99-101: "In contrast to the
extensive knowledge about protein biochemestry of keratinization, little
is known about the fate of nuleic acids during these processes. It has
been suggested that both DNA and RNA completly degenerate in the initial
phase of keratinization. However, from electron microscopial studies it
is known that nuclear remnants are present. From these findings and our
own observations one might expect that nucceic acids should still be
available...Although our study describes preliminary experiments and
must be confirmed by further investigations the persistence of
high-molecular-mass DNA within (fresh fingernail) cells which die in
their normal differentiation may give new aspects to the process of
keratinization."
2. R. Reynolds, et al., The Development and Evaluation of New Genetic
Markers for the Application of PCR to Forensic Casework,Advances in
Forensic Haemogenetics 4 (1991) 29-31: "Frequently forensic DNA
samples are degraded to some extent. It has been demonstrated previously
that amplification of severely degraded DNA samples yields either the
correct HLA DQA type or no type. More recent studies indicate it is
possible to obtain ambigious and even incorrect D1S80 types from
moderately degraded DNA samples."
3.P.T. Bilous, et. Al. DNA Analysis of human Blood Recovered from
Explosion Debris,Advances in Forensic Haemogenetics 5 (1993) 265-267:
"The majority of bloodstains (11 of 15 ) yielded DNA typing
profiles ( single locus RFLP analysis) consistent with the donor control
samples. Partial digestion, altered migration, and weak intensity bands,
were encountered in the remaining four samples. Three of these samples
were found exposed to the soil."
4.G.A. Drago, et. Al., The Degradation of a VNTR Polymorphic Site In
Various Biological Specimens When Subjected To Different Temperature
Conditions,Advances in Forensic Haemogenetics 5 (1993) 271-274:
"These results indicate that DNA may be more labile than previously
observed....DNA from liquid semen shows some degradation even within 1
week. Material extracted from liquid blood and semen stored at 37
degrees C shows weak band patterns after one week and no bands at all
were detected after 4 weeks. The observation in some instances in
material from liquid semen and semen stains, both stored at room
temperature, of bands additional to those expected gives cause for
concern and warrants further investigation. In this limited study
although such bands were seen in association with the bands which were
expected to be present in the sample they could appear stronger than the
expected bands." (RFLP)(Department of Haematology, The london
Hospital Medical College)
5.M. Prinz, et. al., Effect of Degradation on PCR Based DNA
Testing,Advances in Forensic Haemogenetics 5 (1993) 375-378: "
Using DNA polymorphisms with short alleles and the PCR approach it is
possible to acheive results from minimal of DNA and degraded samples.
But, as has been described e.g. by Reynolds et al. (1993) for D1S80, the
degradation can lead to dubious or even wrong typing results through
causing preferential or sole amplification of the smaller allele. In
cases where the degree of degradation cannot be tested beforehand,
degradation artefacts will not always be expected and recognized.
6. P. Wiegand et al., DNA Typing if Debris From Fingernails, Int. J.
Leg. Med. (1993) 106:81-83:" especially important is the method of
preparation. A careful but not extensive cleaning procedure is the most
successful because more extensive procedures lead to an excess of DNA
from the fingernails of the scratcher. This can give rise to allelic
drop-out or to a quantitative suppression of the other
contributor."
7. P. Fattorini et al., False Results in the HLA-DQ alpha Typing: Two
Cases Reported, Advances in Forensic Haemogenetics 6 (1995) 272-274:
" two cases are reported where ambigious results by HLA-DQ alpha
typing...were obtained. The case one refers to the DNA typing of 13
-year -old bloodstains. The reasons of these results are unknown and
several mechanisms could be involved: contamination by iexogenous DNA,
inhibitors, etc."
8. C. Augustin et al., Identification Of Human Remains:Short Tandem
Repeat Analysis Versus Mitochondrial DNA,Advances in Forensic
Haemogenetics 7 (1997) 83-84: Parts of a severely decomposed body were
analyzed at 8 STR loci. "PCR-analysis... led to contradictory
results between the bone and muscle tissue of the corpse. ..(I)t
remained unclear why the bone did not fir to muscle tissue and hair of
the corpse"
9. C. Capelli, et. Al., Fingerprints From Fingernails?,Advances in
Forensic Haemogenetics 7 (1997) 100-102: " although DNA profiles
are currently reproduced from different biological sources ( hair roots,
semen, faeces), keratin cutaneous appendages have been traditionally
regarded as refractory to DNA extraction and genomic analysis....Nails,
as other cutaneous structures, derive from hyperkeratinization of
epidermal cells...The aim of this study is to evaluate nails as a
candidate to forensic analysis and to test several inherent extraction
procedures to optimize the harvest of DNA from this tissue...Short
incubation and pK concentrations ... returned highly degraded DNA from
fragments down to 100 bp. Unexpectedly, these samples were not
amplifiable (at STR loci), by the presumable coextraction of inhibitors
(keratin itself could be one). In this case, inhibitors are possibly not
destroyed, as is with prolonged enzymatic digestion. On the other hand,
incubation with very low pK concentration should let both DNA and
inhibitors be released and the resulting samples still amplifiable. A
tentative explanation for the highly degraded status of DNA harvest
DNA...could be in the apoptosis ( cellular programmed death, associated
to nuclear DNA fragmentation) to which keritinized tissues undergo. To
follow their psysiological fate, nail cells are likely to activitate
specific biochemical pathways, leading to nuclear membrane disintegation
and keritin content increase. In this process, the DNA is set free in
the cytoplasm, it is attacked and psysiologically degraded by
nucleases.)
10. R. Roby, The Extraction of DNA From Human Nail Material,
Proceedings From the Seventh International Symposium on Human
identification (Promega Corp. 1996) p. 133: "In order to process
nails as evidence, a validation study was performed to demonstrate that
DNA could be susuccessfully extracted from human nail material. Organic
extraction of aged (8 year old) human nail material yielded a sufficient
quantity of DNA for successful mtDNA sequencing; however, STR analysis
was unsuccessful." (Study by The Department of Defense DNA
Registry).
This review of the literature indicates persuasively that although
some initial validation work has been done on the typing of human
fingernails, the state of scientific knowledge is not such that typing
of thirteen year old fingernails scrapings is accepted as reliuable even
by forensic scientists, let alone by the broader community of molecleur
biologists. In fact , if the last quoted study by the United States
Department of Defense is to be credited, reliable typing of such aged
stains is not possible.
c.Misincorporation, Differences in Qualitative and Quantitative
Fidelity, and Differential Amplification.
PCR amplification products do not always faithfully represent the
starting material in the sample, either qualitatively or quantitatively.
During PCR amplification nucleotides are known to be "misincorporated"
at the rate of one per 10,000 nucleotides per cycle. If amplification is
performed on a sample that has a large number of molecules, and if the
misincorporation is random ("stochastic fluctuation"), then
the low frequency of random errors will not skew typing results. The NRC
I warns, however, that "for systems in which misincorporation is
not random," difficulties will arise. NRC I, at 64. In particular,
in DNA systems that contain tandem repeat sequences, "the DNA
polymerase can slip during amplification, introduce or delete copies of
the repeat, and produce a heterogeneous collection of fragments, often
making interpretation difficult." Id. Recent literature shows that
the misincorporation problem has been the most serious for STRs and
D1S80 (See, section III A. 2.c.iii, # 1-4, 6, 16, 18 infra).
Differential amplification is also a potential problem:
In some cases, PCR can be qualitatively faithful but quantitatively
unfaithful, because some alleles amplify more efficiently than others. A
sample might contain a 50:50 mixture of two alleles and yield an
amplified product with a 90:10 ratio. Differential amplification can
arise through several mechanisms. It has been observed in the
amplification of allelic products of different sizes (larger products
tend to amplify less efficiently than shorter products) and in the
amplification of sequences that differ significantly in GC content
(because of differing denaturation efficiencies). In some cases,
faithful amplification occurs at some temperatures and differential
amplification at other temperatures. The possibility of differential
amplification needs to be addressed in the design and development of
amplification protocols for each genetic marker system. The safeguards
to ensure that differential amplification does not occur should be
defined and documented.
NRC 1 at 64-65.
When dealing with mixed forensic samples, especially small degraded
samples of the kind at issue in this case, quantitative analysis with
current forensic PCR methods can be, in the NRC's words,
"problematic." Id. at 65. The example selected by the NRC I as
representative of this problem is the exact dot blot intensity analysis
used by Mr. Keel in this case. The NRC I also notes that " (i)deally,
primers should amplify only the desired target locus. However,
nonspecific amplification can be seen, if one amplifies for extended
cycle numbers. Limits on the cycle number might be required as a
safeguard against nonspecific products" Id. at 65. The evidence in
Bokin indicates clearly that Mr. Keel does not monitor his cycling
parameters. (See, Exhibit E at 5)
It should be noted that Dr. Mullis believes that PCR based sequencing
methods, such as those being developed now on mitochondrial DNA, can
avoid the problems of quantitative and qualitative amplification:
But the sequencing -- the reason I was so strong on sequencing is that
it doesn't rely -- sequencing doesn't rely on the quantitative accuracy
in the amplification of different bands. The quantitative aspect is not
as important. Just as long as you've got enough to sequence, it won't
matter if you were sort of inefficient in your amplification if you have
enough to sequence. If you don't, you can just amplify some more until
you do...
If you're sequencing -- if you're sequencing with PCR amplification,
[inefficient amplifications] won't change the results. It will say well,
you didn't get enough to even get a sequence, so go back and amplify
some more. It won't change the sequence. The data that you come out with
will be the same, regardless of whether any of those things are wrong.
McIntosh, Mullis, 2377 line 13 to 2378 line 13.
d. False Positives
i. Introduction In their original points and authorities in support of
admission of DNA evidence in both Bokin and McClanahan, the People
provided the court with an impressive looking Appendix containing 95
citations to articles in forensic science trade journals.This list of
titles was presented for the apparent purpose of demonstrating an
alleged universal acceptance of all forms of PCR-based testing in every
aspect of science and medicine. The court was also informed that "(t)he
People are not aware of any publications questioning the scientific
reliabilty of PCR-based DNA analysis and population databases." (
Bokin Points and Authorities at p. 76 )
The People's claimed ignorance is reminiscent of the prosecution's two
Kelly-Frye experts in People v. Morganti (1996) 43 Cal. App. 3rd 656,
both of whom testified in April, 1993 that they were "not aware of
any scientists who work in this field or of any publications expressing
the view that PCR analysis is unreliable." Id. P. 664; See also,
People v. Wright (1998) 62 Cal. App. 4th 31 (one of the same experts, Ed
Blake, provided the same testimony in August, 1994 )
A close look at the articles cited by the People reveals that many of
them do not consider the specific methods and procedures used by the San
Francisco Crime Lab in this case. Moreover, most, if not all, of the
articles have been authored by law enforcement personnal whose
livelihood is dependent on court acceptance of PCR or by employees of
companies who stand to make huge profits from such acceptance. As noted
in the National Research Council's first report on DNA Technology in
Forensic Science (NRC I ): " There is reason to be skeptical of
entrusting any important regulatory matters to a self-regulating
organization.... There has been a notable dearth of published research
in forensic DNA testing by scientists unconnected to the companies that
market the tests." (p.162)
More fundamentally, it was not true when Morganti and Wright were
litigated, and it is certainly not true today, that there are no peer
reviewed publications, by disinterested scientists, questioning the
scientific reliabilty of PCR-based DNA analysis and population
databases. The People's pleading in Bokin made the claim that the
defense had not cited any scientific literature which post-dates the
Wright appellate decision in 1998. But since Wright, contrary to Soto,
explicitly refused to consider any testimony by defense experts that was
elicited after the Kelly hearing in that case ( 62 Cal. App. 4th at
41-42 ), and Morganti expressly but erroneously holds that " (s)ince...
a court should ordinarily confine itself to the record on matters of a
highly complex nature, we have not consulted outside literature in this
case " ( 43 Cal. App. 4th at 663 n.10) , the defense urges that the
dates of the Kelly hearings in those cases should be the proper focus of
the inquiry as to whether " the prevailing scientific opinion has
materially changed". People v. Venegas (1998) 18 Cal. 4th 47 In
what follows, the defendant will attempt to provide a summary of only
some of the most recent general scientific and forensic literature which
calls into serious question the reliabilty of PCR-based DNA testing in
medical science in general and forensic science in particular.
Responsive to the People's pleadings in Bokin, the focus here is on the
more recent literature in order to demonstrate that to the extent they
are relevent at all, Morganti and Wright should not be controlling in
this case because the prevailing scientific opinion has materially
changed since those cases were litigated in April 1993 and August 1994.
ii.Recent Scientific Literature Demonstrates That Prevailing
Scientific Opinion Has Materially Changed With Respect To The
Reliability Of PCR-Based DNA Testing In Medical Science
A recent MEDLINE search of scientific articles on "false
positive PCR tests" revealed a total of 315 general nonforensic
medical science articles in the period 1993 to the present, as compared
to only 82 such articles for the period 1989 -1992. Abstracts of only
some of the most recent articles are as follows:
1. 7. Borrow R; Guiver M; Sadler F; Kaczmarski EB; Fox AJ False positive
diagnosis of meningococcal infection by the IS1106 PCR ELISA FEMS
Microbiol Lett, 162(2):215-8 1998 May 15 : At a time when optimal case
ascertainment for meningococcal infection is a high priority, the need
for non-culture case confirmation, in particular by DNA amplification,
is seen as being of vital importance to assist contact management and
cluster recognition. A solution hybridisation assay with colorimetric
microtitre plate detection (polymerase chain reaction-enzyme-linked
immunosorbent assay (PCR ELISA) has been developed using the multicopy
insertion sequence IS1106 which had reportedly achieved a specificity of
100% and was described as being meningococcal specific. This PCR ELISA
assay was evaluated on specimens from over 5000 patients at the national
Meningococcal Reference Unit (MRU) between late 1995 and early 1997 and
was found to be highly sensitive. Insertion sequences, however, are
genetically mobile with the ability to spread between species and even
genera. During the evaluation period of the IS1106 PCR ELISA a number of
false positives proved to be caused by organisms other than N.
meningitidis were recorded resulting in the withdrawal of this assay as
a front line screening assay for routine confirmation of meningococcal
infection.
2. Sakugawa H; Nakasone H; Nakayoshi T; Kinjo F; Saito A; Yakabi S;
Zukeran H; Miyagi Y; Taira R; Koja K; et al High proportion of false
positive reactions among donors with anti-HCV antibodies in a low
prevalence area J Med Virol, 46(4):334-8 1995 Aug :Among 39,656
voluntary blood donors in Okinawa Prefecture, Japan, 115 (0.29%) were
repeatedly reactive for antibody to hepatitis C virus (anti-HCV) by
second generation (2nd-gen) passive hemagglutination assay (PHA).
Positive serum samples were tested for anti-HCV using three different
enzyme immunosorbent assays (ELISAs; Abbott 2nd EIA, UBI-HCV-EIA, JCC-2)
and for HCV-RNA by the polymerase chain reaction (PCR). The 115 2nd-gen
PHA-positive sera were divided into three groups according to the
agglutination titers; > 2(10) (high titer group), 2(7)-2(9) (median),
2(5)-2(6) (low). All but one serum (44/45) in the high PHA titer group
reacted in each of the three second screening ELISAs. Furthermore, 43
(97.7%) of the 44 sera contained HCV-RNA by PCR. In the median titer
group, 11 of the 13 samples tested were positive by each of the three
ELISAs, and 4 (36.4%) of the 11 showed reaction by PCR. On the other
hand, all of the 38 sera tested in the low titer group were negative for
HCV-RNA by PCR, and 24 of the 38 were also negative by each of the three
ELISAs. Most of the low titer positive reactions in the 2nd-gen
agglutination assay seemed to be false positive.
3. Kaul K; Luke S; McGurn C; Snowden N; Monti C; Fry WA Amplification of
residual DNA sequences in sterile bronchoscopes leading to false-positivePCR
results J Clin Microbiol, 34(8):1949-51 1996 Aug : PCR has been used
successfully for the direct detection of Mycobacterium tuberculosis in
uncultured patient samples. Its potential is hindered by the risk of
false-positive results as a result of either amplicon carryover of
cross-contamination between patient samples. In the present study, we
investigated whether residual amplifiable human or M. tuberculosis DNA
could remain in sterile bronchoscopes and potentially be a cause of
false-positive PCR results in subsequent patient samples. Sterilized
bronchoscopes were flushed with sterile saline, and the collected eluate
was submitted for PCR amplification of IS6110 sequences and exon 8 of
the human p53 gene. Of a total of 55 washes of sterile bronchoscopes
from two institutions, 2 (3.6%) contained amplifiable M. tuberculosis
DNA and 11 (20%) contained residual human DNA. These findings indicate
that residual DNA can remain in sterilized bronchoscopes and can be a
source of false-positive PCR results.
4. Keum WK; Park CE; Lee JH; Khil LY; Kang I; Kim SS; Jung JC; Oh SM;
Woo HJ; Lee JH; Kim YC; Yoon Y; Choi JW; Ha J Primers determine the
sensitivity of PCR-mediated hepatitis B virus DNA detection and
pretreatment of PCR mixture with 8-methoxypsoralen eliminates
false-positive results Mol Cells, 7(2):244-50 1997 Apr 30: Most methods
for the diagnosis of hepatitis B virus (HBV) infection largely depend on
viral DNA detection by polymerase chain reaction (PCR) or
radioimmunological assay of viral antigens or antibodies. The quality
assurance program recently established in Europe reported that PCR-mediated
HBV DNA detection methods used in many laboratories produced a high rate
of false-positive and false-negative results. Thus, we attempted to
improve the conditions of current PCR methods for detection of HBV DNA.
In the present study, we applied a recently developed method of
releasing HBV DNA from virion by NaOH treatment of patient serum. Using
four different primer sets specific to the HBV core region, we found
that the sensitivity of first-round PCR can be improved by more than two
orders of magnitude depending on the primers. The second round of PCR
using nested primers was sensitive enough to detect up to 10(-6) pg of
the HBV DNA, which is equivalent to approximately 3 copies of the HBV
genome. Among the approximately 800 HBV-infected patient sera
investigated in our laboratory, more than 60% of the tested samples gave
positive results in the first-round PCR. The rate of positive results
obtained using our experimental conditions is very high in comparison
with other reports. The reamplification of the first-round PCR reaction
mixture with the nested primers produced practically 100% positive
results. For diagnosis of HBV infection, we routinely used 1 microliter
of patient serum, which was found to be optimum in our laboratory.
Surprisingly, from 20% of our positive results, even serum diluted to
1/100 (0.01 microliter) produced a stronger signal than 1 microliter.
This observation suggests that direct PCR amplification of HBV DNA
released from serum by NaOH treatment has to be compensated by other DNA
detection methods for correct quantitation. In order to eliminate the
false positive signal resulting from the carry-over due to massive
screening of a large number of samples, PCR reaction mixture containing
8-methoxypsoralen was exposed to ultraviolet light prior to thermal
cycle amplification. This exercise did not decrease the sensitivity of
the detection method, but almost completely removed the false positive
results caused by contaminated templates. We are in the process of
improving PCR-mediated HBV DNA detection methods to attain more reliable
and easily applicable methods.
5. Teo IA; Shaunak S Polymerase chain reaction in situ: an appraisal of
an emerging technique Histochem J, 27(9):647-59 1995 Sep : Polymerase
chain reaction (PCR) in situ is a new technique which promises to
enhance considerably our ability to detect a few copies of target
nucleic acid sequences in fixed tissues and cells. It has an enormous
potential for application in diagnostic histopathology of viral diseases
and in the study of gene expression. PCR in situ is, however,
technically difficult, and amplification of the target DNA is only
30-300 fold. In this article we present an overview of PCR in situ
techniques used to amplify both DNA and RNA targets (RT-PCR in situ). We
also identify problems which can reduce the efficiency of the technique
or which can give rise to false-positive results. They include (1) the
inhibitory effects of cross-linking of histones to DNA or PCR
amplification, (2) abstraction of PCR reagents by tissue-bonding agents
which are used to coat glass slides, (3) poor denaturation of target DNA
and subsequent DNA renaturation due to extensive cross-linking of
histones to DNA, or because of incorrect temperature regulation of
thermal cyclers, (4) false-positive results which arise from end-labelling
of DNA strand breaks by Taq polymerase, and (5) diffusion of PCR
products into and out of cells leading to false-positive results. We
present some of the approaches that have been used to overcome some of
these difficulties and suggest new avenues for investigation to improve
this technique further.
6. Teo IA; Shaunak S PCR in situ: aspects which reduce amplification and
generate false-positive results Histochem J, 27(9):660-9 1995 Sep : PCR
in situ promises the ability to amplify and detect very low levels of
target nucleic acid in tissues. Despite considerable effort, the
technique is still technically difficult and has not yet proved to be
reliable or reproducible. We have now identified a number of factors
which can contribute to the poor amplification of the target DNA and to
the generation of false-positive signals. These factors include the
effects of fixation, reagent abstraction, DNA degradation, DNA end-labelling
and product diffusion. We present evidence to show that formaldehyde
fixation cross-links histones to DNA and thus restricts the subsequent
amplification of target sequences by PCR. End-labelling of DNA occurs
when direct incorporation is used to detect amplified products and this
gives rise to false-positive signals. Amplified products can also
diffuse out of cells and into neighbouring cells which do not contain
target sequences. They can undergo re-amplification within these cells
giving rise to false-positive signals. We believe considerable caution
should be exercised in the interpretation of results generated using PCR
in situ.
7. Boyd AS; Annarella M; Rapini RP; Adler-Storthz K; Duvic M.
False-positive polymerase chain reaction results for human
papillomavirus in lichen planus. Potential laboratory pitfalls of this
procedure J Am Acad Dermatol, 35(1):42-6 1996 Jul :BACKGROUND: The
polymerase chain reaction (PCR) is a common tool in laboratory
evaluation of tissue samples. It has been particularly useful in testing
for nucleic acids of infectious organisms implicated in the pathogenesis
of cutaneous diseases. OBJECTIVE: Our purpose was to evaluate archival
biopsy material (paraffin-embedded) of lichen planus for human
papillomavirus (HPV) DNA with PCR and in situ hybridization (ISH).
METHODS: Tissue sections were obtained from 13 specimens of lichen
planus and 10 control tissues. Standard methods for PCR and ISH were
used to evaluate the presence of HPV DNA. RESULTS. Initial PCR results
demonstrated HPV DNA in 11 of 13 specimens and in none of the control
tissue. However, testing for specific HPV types revealed nucleic acid
contamination. ISH was negative in all specimens. CONCLUSION: Accurate
evaluation of tissue with PCR is difficult because of the procedure's
profound sensitivity. Positive results reported in the literature should
be viewed with caution. Potential causes for false-positive and
false-negative results should be considered.
8. Wiedbrauk DL; Werner JC; Drevon AM Inhibition of PCR by aqueous and
vitreous fluids.J Clin Microbiol, 33(10):2643-6 1995 Oct : The detection
of viral nucleic acids in intraocular fluids and tissues by PCR has
become increasingly important in clinical ophthalmology. While much
attention has been directed toward minimizing false-positive reactions
resulting from specimen contamination or amplicor carryover, relatively
little attention has been given to the causes of false-negative PCRs.
This report describes a PCR inhibitor in normal aqueous and vitreous
fluids that can produce false-negative PCR results. As little as 0.5
microliter of vitreous fluid and 20 microliters of aqueous fluid can
completely inhibit DNA amplification in a 100-microliters PCR mixture.
This inhibition was not primer specific, nor was it due to chelation of
Mg2+ ions or DNase activity in the ocular fluid. The inhibitor was
completely resistant to boiling for 15 min. However, the inhibitory
effects were completely removed by a single chloroform-isoamyl alcohol
(24:1) extraction. The extent of PCR inhibition depended upon the type
of thermostable DNA polymerase used in the reaction. Taq DNA polymerase
was very sensitive to the inhibitor, while thermostable DNA polymerases
from Thermus thermophilus HB-8 (Tth) and Thermus flavus (Tfl) were
completely resistant. Thus, the inhibitory effects of intraocular fluids
on PCRs can be removed by diluting the specimen, by chloroform
extraction, or by using Tth or Tfl DNA polymerases
9. Smith KC; Starke JR; Eisenach K; Ong LT; Denby M Detection of
Mycobacterium tuberculosis in clinical specimens from children using a
polymerase chain reaction .Pediatrics, 97(2):155-60 1996 Feb :
OBJECTIVE: We evaluated the usefulness of the polymerase chain reaction
(PCR) using the insertion sequence IS6110 as the target for DNA to
detect Mycobacterium tuberculosis in clinical specimens from children.
STUDY DESIGN: This was a prospective, controlled, blinded study
comparing PCR on clinical specimens, mycobacterial culture, and clinical
diagnosis. PATIENTS: Sixty-five hospitalized children were evaluated, 35
with tuberculosis disease and 30 controls. Cases were defined by culture
and/or specific clinical criteria. Controls included patients with
tuberculosis infection but no detectable disease as well as patients
free of tuberculosis infection and disease. RESULTS: Polymerase chain
reaction had a sensitivity of 40% and a specificity of 80% compared with
clinical diagnosis. Mycobacterial culture had a sensitivity of 37%. The
combination of culture and PCR identified 19 of 35 children (54%) with
clinically diagnosed tuberculosis. There were six children with
false-positive PCR results: One had tuberculosis infection without
disease, two had Mycobacterium avium lymphadenitis, and three had
diagnoses unrelated to tuberculosis. CONCLUSIONS: The sensitivity of PCR
is comparable to that of culture for detecting M tuberculosis in
children, and may strengthen and hasten the clinical diagnosis in
culture-negative patients. However, because of the limitations in
specificity, the results of PCR alone are insufficient to diagnose
tuberculosis in children. Although ongoing refinements in PCR techniques
should improve the specificity of this test, epidemiologic and clinical
information continue to be the most important consideration in the
diagnosis of tuberculosis in culture-negative children.
10.Bahrmand AR; Bakayev VV; Babaei MH Use of polymerase chain reaction
for primary diagnosis of pulmonary tuberculosis in the clinical
laboratory. Scand J Infect Dis, 1996, 28:5, 469-72: A nested polymerase
chain reaction (PCR) has been used for the rapid detection of tubercule
bacilli in respiratory specimens from 287 patients suspected of
tuberculosis. The results of PCR testing were compared with isolation
methods (conventional culture and Bactec system) in 110 smear-positive
and 177 smear-negative patients. There were only 4 false negative
results by PCR in the 171 specimens that were M. tuberculosis complex
culture-positive. Of 92 PCR-positive samples prepared from the
smear-positive specimens 90 (97.8%) were confirmed by culture. However,
a poor correlation was obtained between initial 122 PCR-positive results
and combined 81 culture recovered organisms in smear-negative patients.
After verification of the efficacy of isolation method, retesting PCR-positive
culture-negative samples, and studies of patients' clinical histories,
only 18 of the cases were found to be associated with the disease. The
other 29 results out of the original 47 discrepants were considered PCR
false positives, possibly due to contamination. In conclusion, the PCR
assay described is suitable for implementation in daily routine work
with respiratory specimens, however it should be validated with culture,
especially for the smear-negative patients.
11.Schwartz DH, Laeyendecker OB, Arango-Jaramillo S, Castillo RC,
Reynolds MJ Extensive evaluation of a seronegative participant in an
HIV-1 vaccine trial as a result of false-positive PCR. Lancet 1997 Jul
26;350(9073):256-259: BACKGROUND: In the USA, more than 2000 volunteers
have received one or more experimental HIV-1 vaccines in phase I and II
clinical trials, and there have been breakthrough HIV-1 infections among
participants receiving vaccine and placebo. Serological diagnosis of new
HIV-1 infections in vaccine-trial participants will become increasingly
complicated as more viral components are used in vaccines. Recognition
of this problem has led to a reliance on viral-genome measurement to
distinguish vaccine-induced immunity from HIV-1 infection. Currently,
quantitative RNA measurement is expensive, prone to contamination, and
reliable only in laboratories certified by manufacturers or that have
quality-control programmes. METHODS: A high-risk vaccinee presented
after an acute febrile episode and was tested for HIV-1 infection by
reverse transcriptase (RT) PCR of viral RNA. Further extensive tests
were required to clarify the HIV-1 infection and immune status of the
vaccinee, including repeat RT-PCR, nested DNA PCR, western blot,
lymphoproliferation assay, cytotoxic T-cell lysis, CD8-depleted
co-culture, and HIV-1 challenge culture. FINDINGS: Initial testing of
plasma by RNA RT-PCR was reported as positive. This result was not
confirmed by viral cultures, nested DNA PCR, western blot, or EIA.
Additional RNA RT-PCR assays gave positive results from earlier
occasions in the vaccine trial. Eventually, testing of all previously
reactive samples by RNA RT-PCR was repeated in a quality-controlled
laboratory, and confirmed the negative HIV-1 status of the individual.
INTERPRETATION: This case report exemplifies the difficulties with use
of viral-genome measurement as a screening test to diagnose HIV-1
infection, particularly in individuals who have ever participated in
HIV-1 vaccine trials. Monitoring of large numbers of phase III vaccinees
by RNA RT-PCR will not be feasible. The design of efficacy trials for
new vaccines should be in parallel with development of antibody-based
diagnostic tests that are capable of differentiating between
immunisation and true HIV-1 infection.
12.Busch MP, Henrard DR, Hewlett IK, Mehaffey WF, Epstein JS, Allain JP,
Lee TH, Mosley JW Poor sensitivity, specificity, and reproducibility of
detection of HIV-1 DNA in serum by polymerase chain reaction. The
Transfusion Safety Study Group. J Acquir Immune Defic Syndr
1992;5(9):872-877 A series of recent studies have reported detection by
the polymerase chain reaction (PCR) of cell-free human immunodeficiency
virus type 1 (HIV-1) DNA (as opposed to virion RNA) in serum from both
seropositive and seronegative persons. To evaluate the sensitivity,
specificity, and reproducibility of PCR detection of cell-free HIV-1
DNA, we distributed coded panels containing 98 serum specimens obtained
from well-characterized, infected individuals and control blood donors
to the two laboratories with reported experience with this technique.
Positive results were reported with HIV-1 gag primers (SK38/39) for 48
of 188 separate PCR determinations on DNA extracts from 44 serum samples
from seropositive patients (25.5% sensitivity). HIV-1 gag signal was
also reported for 28 of 151 PCR determinations on 34 samples from
noninfected blood donors (18.5% false-positive rate). PCR for HIV-1 env
DNA performed in one laboratory was negative on all specimens from
seropositive and seronegative patients. Results for cell-free HIV-1 gag
and human genomic (beta-globin or HLA DQ-alpha) DNA were inconsistent on
replicate and serial specimens evaluated within each laboratory and
between laboratories. These results indicate that current techniques for
detecting cell-free HIV-1 DNA in serum lack adequate sensitivity,
specificity, and reproducibility for widespread clinical applications.
13.McHugh TD, Ramsay AR, James EA, Mognie R, Gillespie SH Pitfalls of
PCR: misdiagnosis of cerebral nocardia infection. Lancet 1995 Nov
25;346(8987):1436: "Histology...did not show features suggestive of
T gondii infection. Pcr resulted in a PCR product similar to that from
the T gonii control.....Our study identifies an unrelated but clinically
significant organism as a source of crossreactivity and confirms that
before P(CR)30 assay can be applied to patient diagnosis, DNA from
organisms likeky to be found in clinical specimens must be examined for
similar false-positive results."
14.Holliman RE, Patel B, Johnson JD, Mangan J, Savva D Pitfalls of PCR:
misdiagnosis of cerebral nocardia infection. Lancet 1996 Feb
3;347(8997):335-336 : " Our results confirm the generation of a PCR
product when applying the P30 based assay to some nocardia stains. We
also found cross-reactivity with Mycobacterium tuberculosis....
Established good practice dictates that PCR should not be the sole
diagnostic test for toxoplasmosis in any clinical setting."
15.Trinker M, Hofler G, Sill H False-positive diagnosis of tuberculosis
with PCR. Lancet 1996 Nov 16;348(9038):1388 :" This case
demonstrates that the results of pcr assays-although highly specific for
the detection of mycobacterial DNA-should be interpreted only in
conjunction with clinical information in order to avoid inappropriate
treatment. PCP assays are not standardized;false positive results may be
due to individual laboratory procedures, contamination, or might reflect
amplification of dna from nonviable myco bacteria."
16.Chen G, Chen Z Effects of reagents used in polymerase chain reaction
on sensitivity and specificity of detection of Mycobacterium
tuberculosis. Chung Hua Chieh Ho Ho Hu Hsi Tsa Chih 1996
Oct;19(5):294-296: OBJECTIVE: To explore the effects of reagents used in
polymerase chain reaction on sensitivity and specificity of detection of
mycobacterium tuberculosis. METHODS: 205 tuberculosis and 105
nontuberculosis cases' sputum specimens were examined simultaneously by
smear, polymerase chain reaction using three different reagents and
culture. RESULTS: The positive rates of Mycobacterium tuberculosis by
polymerase chain reaction using three different reagents were 82.4%,
71.7% and 61.0% respectively in 205 tuberculosis cases, while 18.0%,
11.4% and 2.8% respectively in 105 nontuberculosis cases. The
differences between the above three different groups was statistically
significant (P < 0.005). CONCLUSIONS: The sequences of primers and
the length of amplified products of different reagents may be associated
with the positive and the false positive rates of Mycobacterium
tuberculosis in polymearase chain reaction.
17.Doucet-Populaire F, Lalande V, Carpentier E, Bourgoin A, Dailloux M,
Bollet C, Vachee A, Moinard D, Texier-Maugein J, Carbonnelle B, Grosset
J A blind study of the polymerase chain reaction for the detection of
Mycobacterium tuberculosis DNA. Azay Mycobacteria Study Group. Tuber
Lung Dis 1996 Aug; 77(4):358-362:SETTING: Nine French laboratories
routinely involved in mycobacterial work. OBJECTIVE: To assess the
detection of Mycobacterium tuberculosis in experimental samples by
polymerase chain reaction (PCR) using the insertion sequence IS6110 as a
target for deoxyribonucleic acid (DNA) amplification. DESIGN: Nine
laboratories participated in a blind study of the detection of M.
tuberculosis by PCR in 20 coded samples containing either a definite
number of M. tuberculosis complex (positive samples) or environmental
mycobacteria (four samples) or no mycobacteria (five samples). RESULTS:
Five laboratories reported false-positive PCR results, with an average
rate of 7%. All laboratories except one reported positive PCR results
for samples containing 10(5) cfu/ml or more. M. tuberculosis DNA was
detected in two thirds of samples containing 10(4) and 10(3) cfu/ml, and
in one third of the samples containing 10(2) cfu/ml. CONCLUSION: The
results of the study suggest that PCR using IS6110 as a target for DNA
amplication is neither very sensitive nor really specific for the
detection of M. tuberculosis.
18.Noordhoek GT, Kolk AH, Bjune G, Catty D, Dale JW, Fine PE, Godfrey-Faussett
P, Cho SN, Shinnick T, Svenson SB, et al Sensitivity and specificity of
PCR for detection of Mycobacterium tuberculosis: a blind comparison
study among seven laboratories. Clin Microbiol 1994 Feb;32(2):277-284 :PCR
is, in principle, a simple and rapid test for use in the detection of
Mycobacterium tuberculosis. However, virtually no data are available on
the reliability and reproducibility of the method. In order to assess
the validity of PCR for the detection of mycobacteria in clinical
samples, seven laboratories participated in a blinded study of 200
sputum, saliva, and water samples containing either known numbers of
Mycobacterium bovis BCG cells or no added organisms. Each laboratory
used its own protocol for pretreatment, DNA extraction, and detection of
the amplification product. Insertion sequence IS6110 was the target for
DNA amplification. Several participating laboratories reported high
levels of false-positive PCR results, with rates ranging from 3 to 20%
and with one extreme value of 77%. The levels of sensitivity also ranged
widely among the different participants. A positive PCR result was
reported for 2 to 90% of the samples with 10(3) mycobacteria. Although
most participants did include control tests to check the sensitivity and
specificity of the PCR, the sequence of operations from sample
pretreatment to purification of DNA from bacteria was not always
monitored adequately. During these procedures cross-contaminating DNA
was introduced and/or bacterial DNA was lost. The results of the study
show that the implementation of an effective system for monitoring
sensitivity and specificity is required before the PCR can be used
reliably in the diagnosis of tuberculosis.
19.du Toit R, Victor TC, van Helden PD Empirical evaluation of
conditions influencing the polymerase chain reaction: enterotoxigenic
Escherichia coli as a test case. Eur J Clin Chem Clin Biochem 1993
Apr31(4):225-231: False polymerase chain reaction (PCR) results may be
obtained under unfavourable reaction conditions. Therefore optimal
conditions for the different factors influencing a specific PCR method
should be determined before introduction to a clinical diagnostic
laboratory. This study has concentrated on the detection of heat-labile
enterotoxin-producing E. coli by PCR, with empirical evaluation of
various factors. Template was prepared by heat-lysis of E. coli, and
this was shown to be adequate for PCR detection. The results showed that
deviation from the optimal conditions of any of the following conditions
may lead to false results: lysis of bacterial cells, denaturation
temperature during cycling, annealing temperature, primer concentration,
enzyme concentration, magnesium concentration and ion concentration.
Three different detection methods for PCR product were also evaluated.
As little as one bacterium can be detected after 35 cycles of PCR
amplification with 32P-labelled oligonucleotide probe. An alkaline
phosphatase-labelled probe was 10-fold less sensitive, whereas 100
bacteria in 10 microliters of the original sample suspension were
necessary to give a positive signal after gel electrophoresis. The
information in this study may be useful to those who wish to introduce
PCR tests to diagnostic laboratories
20.Wood GS, Salvekar A, Schaffer J, Crooks CF, Henghold W, Fivenson DP,
Kim YH, Smoller BR Evidence against a role for human T-cell
lymphotrophic virus type I (HTLV-I) in the pathogenesis of American
cutaneous T-cell lymphoma. J Invest Dermatol 1996 Sep;107(3):301-307 :We
used a standard polymerase chain reaction (PCR)/Southern blot assay
(sensitivity > 10(-5)) to detect human T-cell lymphotrophic virus
type I (HTLV-I) proviral pX, pol, and env genes in the lesional skin of
42 American patients with cutaneous T-cell lymphoma (CTCL). As in some
prior reports using similar methods, a variable proportion of PCR tests
were positive (seven of 42 for pX, three of 42 for pol, and two of 37
for env), resulting in an overall positive test rate of 12 of 121 (10%).
To determine the significance of these positive test results, we
performed several additional studies. D1S80 polymorphism analysis of
CTCL cases and HTLV-I PCR analysis of non-CTCL dermatosis controls
showed no evidence that positive PCR tests resulted from sample
mislabeling, gross HTLV-I contamination, or human endogenous
retroviruses. We then modified the standard PCR assay to incorporate
ultraviolet (UV) light to destroy low-level PCR contamination. With this
modified assay (sensitivity > 10(-5)), only three of 12 previously
positive cases were still positive, suggesting that the earlier
positives were due to trace contamination of PCR reagents or trace
contamination of sample DNA. This interpretation was also supported by:
(i) a match between pX and pol sequences cloned from one PCR-positive
specimen and the MT4-positive control, (ii) our inability to confirm
HTLV-I in any PCR-positive case using genomic dot blotting (sensitivity
> 10(-2)), and (iii) negative PCR results when new samples from two
of the remaining positive cases were analyzed. Finally, we used our
modified UV/ PCR/Southern blot assay to test an additional 28 cases of
American CTCL for pX. All of them were negative. Although these studies
of 70 cases of American CTCL do not exclude the possibility that another
virus is involved in the pathogenesis of this disease, they provide
strong evidence against a role for HTLV-I. Furthermore, they emphasize
the need for special strategies to control for false-positive PCR tests
that can result from even trace levels of contamination with viral DNA.
As a consequence, associations between diseases and viruses should be
viewed skeptically if they are based primarily on conventional PCR data.
21.Wilke WW, Jones RN, Sutton LD Department of Pathology, University of
Iowa College of Medicine USA. Automation of polymerase chain reaction
tests. Reduction of human errors leading to contamination. Diagn
Microbiol Infect Dis 1995 Apr;21(4):181-185: We compared the performance
over 21 months of manually performed polymerase chain reaction (PCR)-based
DNA analysis experiments with 25 months of automated PCR performed by a
Zymark robotic system. Automation of the PCR technique resulted in a
sixfold reduction in the number of experiments reporting carryover
contamination and decreased the overall rate of contamination among
total reactions 68-fold. Whereas contamination occurrences among manual
experiments were evenly dispersed over the study interval and correlated
with the lack of experience of laboratory personnel, the contamination
that occurred with the robotic system was confined to the first 10
months of operation. In manual experiments, many of the 81 no-target
false positives were sufficiently strong to result in the invalidation
of 151 samples and positive controls. The seven no-target control false
positives in the automated system were weak bands that were easily
subtracted as background. Because none of the negative samples had DNA
bands, no sample on the automated system has ever been invalidated as a
result of contamination. Automation of PCR tests appears to offer great
promise in reducing contamination to acceptable levels
22.Kelley PK, McClain KL Department of Pediatrics, Baylor College of
Medicine, Respiratory contamination of polymerase chain reactions by
human herpesvirus 6. Am J Hematol 1994 Dec;47(4):325-327: Human
herpesvirus type 6 DNA derived from human breath was discovered to
contaminate PCR reactions during routine reaction preparation. Parallel
PCR experiments were conducted in which expiratory secretions were
blocked by a surgical mask, while others were performed without any
attempt to circumvent respiratory contamination. The experimenter was
previously determined to harbor HHV-6 DNA in the saliva. All reactions
in which expiration was obstructed were negative for HHV-6 DNA via PCR.
Reactions in which there was no attempt to obstruct respiratory
secretions were positive for HHV-6 DNA. These data suggest that PCR
assays investigating the presence of HHV-6 may be highly susceptible to
contamination from the experimenter leading to false positive results.
iii. Recent Scientific Literature Demonstrates That Prevailing
Scientific Opinion Has Materially Changed With Respect To The Reliabilty
Of PCR-Based DNA Testing In Forensic Science
Among the more recent of the scientific articles on the forensic
application of PCR-based DNA testing are the following:
1.Smith JR, Carpten JD, Brownstein MJ, Ghosh S, Magnuson VL, Gilbert DA,
Trent JM, Collins FS National Center for Human Genome Research, National
Institutes of Health, Approach to genotyping errors caused by
nontemplated nucleotide addition by Taq DNA polymerase. Genome Res 1995
Oct;5(3):312-317: Thermostable DNA polymerases can catalyze nontemplated
addition of a nucleotide to the 3' end of amplification products. This
presents a potential source of error in genotyping studies employing Taq
DNA polymerase to amplify microsatellite loci. Although the activity is
marker specific, experimental variation is often seen in the degree of
modification. Consequently, for a given microsatellite marker, an allele
may be inconsistently identified as either the unmodified or modified
amplification product. Full automation of high-throughput genotyping has
been hampered by the need for manual editing of data because of this
source of allele misidentification. In this study we estimate a 1% to 3%
error rate attributable to nontemplated nucleotide addition in the ABI
PRISM genotyping system. We present a PCR-based strategy to minimize
this source of error.
2.. Brownstein MJ, Carpten JD, Smith JR National Institute of Mental
Health, National Institutes of Health Biotechniques 1996
Jun;20(6):1004-1006 Modulation of non-templated nucleotide addition by
Taq DNA polymerase: primer modifications that facilitate genotyping. Taq
DNA polymerase can catalyze non-templated addition of a nucleotide
(principally adenosine) to the 3' end of PCR-amplified products.
Recently, we showed that this activity, which is primer-specific,
presents a potential source of error in genotyping studies based on the
use of short tandem repeat (STR) markers. Furthermore, in reviewing our
data, we found that non-templated nucleotide addition adjacent to a 3'
terminal C is favored and that addition adjacent to a 3' terminal A is
not. It was clear, however, that features of the template in addition to
the 3' terminal base also affect the fraction of product adenylated. To
define consensus sequences that promote or inhibit product adenylation,
we transplanted sequences between the 5' ends of the reverse primers of
markers that are adenylated and those of markers that are not adenylated.
It proved difficult to identify a single sequence capable of protecting
the products of all markers from non-templated addition of nucleotide.
On the other hand, placing the sequence GTTTCTT on the 5' end of reverse
primers resulted in nearly 100% adenylation of the 3' end of the forward
strand. This modification or related ones (called
"PIG-tailing") should facilitate accurate genotyping and
efficient T/A cloning.
3. Magnuson VL, Ally DS, Nylund SJ, Karanjawala ZE, Rayman JB, Knapp JI,
Lowe AL, Ghosh S, Collins FS National Center for Human Genome Research,
National Institutes of Health Substrate nucleotide-determined non-templated
addition of adenine by Taq DNA polymerase: implications for PCR-based
genotyping and cloning. Biotechniques 1996 Oct;21(4):700-709: The
Applied Biosystems PRISM fluorescence-based genotyping system as well as
the Invitrogen TA Cloning vector system are influenced by the tendency
of Taq DNA polymerase to add an adenine nucleotide to the 3' end of PCR
products after extension. Incomplete addition of adenine to a majority
of PCR product strands creates problems in allele-calling during
genotyping and potentially diminishes the cloning efficiency of such
products. Experiments reported here show that certain terminal
nucleotides can either inhibit or enhance adenine addition by Taq and
that PCR primer design can be used to modulate this activity. The
methods we propose can substantially improve allele-calling for
problematic microsatellite markers when using GENOTYPER software.
4. Cline J, Braman JC, Hogrefe HH Stratagene Cloning Systems PCR
fidelity of pfu DNA polymerase and other thermostable DNA polymerases.
Nucleic Acids Res 1996 Sep 15;24(18):3546-3551:Replication fidelities of
Pfu, Taq, Vent, Deep Vent and UlTma DNApolymerases were compared using a
PCR-based forward mutation assay. Averageerror rates (mutation
frequency/bp/duplication) increased as follows: Pfu (1.3x 10(-6)) <
Deep Vent (2.7 x 10(-6)) < Vent (2.8 x 10(-6)) < Taq (8.0 x
10(-6))< < exo- Pfu and UlTma (approximately 5 x 10(-5)). Buffer
optimizationexperiments indicated that Pfu fidelity was highest in the
presence of 2-3 mMMgSO4 and 100-300 microM each dNTP and at pH 8.5-9.1.
Under these conditions,the error rate of exo- Pfu was approximately
40-fold higher (5 x 10(-5)) thanthe error rate of Pfu. As the reaction
pH was raised from pH 8 to 9, the errorrate of Pfu decreased
approximately 2-fold, while the error rate of exo- Pfuincreased
approximately 9-fold. An increase in error rate with pH has also
beennoted for the exonuclease-deficient DNA polymerases Taq and exo-
Klenow,suggesting that the parameters which influence replication error
rates may besimilar in pol l- and alpha-like polymerases. Finally, the
fidelity of 'longPCR' DNA polymerase mixtures was examined. The error
rates of a Taq/Pfu DNApolymerase mixture and a Klentaq/Pfu DNA
polymerase mixture were found to beless than the error rate of Taq DNA
polymerase, but approximately 3-4-foldhigher than the error rate of Pfu
DNA polymerase.
5. Hughes MS, Beck LA, Skuce RA Department of Agriculture for Northern
Ireland, Stromont, Belfast. J Clin Microbiol 1994 Aug;32(8):2007-2008
Identification and elimination of DNA sequences in Taq DNA
polymerase.This study confirms that different preparations of Taq DNA
polymerase are contaminated with eubacterial DNA. The contaminants
appeared to represent more than one strain or species but were not
identified as Thermus aquaticus or Escherichia coli. Differences in
microcentrifuge tube composition appeared to affect elimination of the
contaminants.
6. Cline J, Braman JC, Hogrefe HH Stratagene Cloning Systems PCR
fidelity of pfu DNA polymerase and other thermostable DNA polymerases.
Nucleic Acids Res 1996 Sep 15;24(18):3546-3551:Replication fidelities of
Pfu, Taq, Vent, Deep Vent and UlTma DNApolymerases were compared using a
PCR-based forward mutation assay. Averageerror rates (mutation
frequency/bp/duplication) increased as follows: Pfu (1.3x 10(-6)) <
Deep Vent (2.7 x 10(-6)) < Vent (2.8 x 10(-6)) < Taq (8.0 x
10(-6))< < exo- Pfu and UlTma (approximately 5 x 10(-5)). Buffer
optimizationexperiments indicated that Pfu fidelity was highest in the
presence of 2-3 mMMgSO4 and 100-300 microM each dNTP and at pH 8.5-9.1.
Under these conditions,the error rate of exo- Pfu was approximately
40-fold higher (5 x 10(-5)) thanthe error rate of Pfu. As the reaction
pH was raised from pH 8 to 9, the errorrate of Pfu decreased
approximately 2-fold, while the error rate of exo- Pfuincreased
approximately 9-fold. An increase in error rate with pH has also
beennoted for the exonuclease-deficient DNA polymerases Taq and exo-
Klenow,suggesting that the parameters which influence replication error
rates may besimilar in pol l- and alpha-like polymerases. Finally, the
fidelity of 'longPCR' DNA polymerase mixtures was examined. The error
rates of a Taq/Pfu DNApolymerase mixture and a Klentaq/Pfu DNA
polymerase mixture were found to beless than the error rate of Taq DNA
polymerase, but approximately 3-4-foldhigher than the error rate of Pfu
DNA polymerase.
7. Herber B; Herold K DNA typing of human dandruff J Forensic Sci, 1998
May, 43:3, 648-56:Dandruff particles contain a considerable portion of
nucleated cells within the aggregates of nuclei-free corneocytes. The
nuclei could be partially degraded due to epidermal differentiation
processes. To test the suitability of DNA from dandruff for forensic
application, a study of 35 subjects and two crime cases has been carried
out using different STR polymorphisms. In 90% of the samples DNA
genotyping could be easily performed indicating that DNA from dandruff
is at least suitable for STR analysis. DNA quantity per dandruff
particle varies extremely within a range from 0.8 to 16.6 ng DNA for the
tested subjects. Genotyping was also possible for a single dandruff
particle if DNA extraction volume had been reduced. In mixed samples
from dandruff coextracted with bloodstains or semen stains dandruff
genotype was detectable. in some cases. These results demonstrate the
relevance of dandruff in forensic casework either as an additional
sample or as a source of inadvertent contamination.
8.Lorente M, Entrala C, Lorente JA, Alvarez JC, Villanueva E, Budowle B
Dandruff as a potential source of DNA in forensic casework J Forensic
Sci 1998 Jul; 43(4):901-902 :Dandruff is a clinical alteration of the
skin that consists histologically of orthokeratotic clumps with minute
parakeratotic foci found in inflammatory pathologies such as seborrheic
dermatitis and psoriasis. Therefore, some nucleated cells should be
found in dandruff and hence there is a possibility that forensically
typeable DNA could be extracted from dandruff. Because of a particular
case in which we were involved, a study was carried out to determine
whether or not DNA could be extracted from dandruff, and if the two most
widely used extraction techniques (Chelex and organic) would be
applicable. Results show that sufficient quantities of DNA (more than 30
to 40 ng) can be obtained from as little as 1.0 to 1.5 mg of dandruff.
Both methods yield DNA, although the organic procedure seems to yield
more (72.5 ng Chelex vs. 183.3 ng organic). All the DNA samples
extracted were typed correctly for the loci HUMTH01 and HUMvWA.
Therefore, dandruff can be considered a potential source of DNA for
forensic identification.
9. Toledano T; Quarino L; Leung S; Buffolino P; Baum H; Shaler RC, An
assessment of DNA contamination risks in New York City Medical Examiner
facilities J Forensic Sci, 1997 Jul, 42:4, 721-4: DNA evidence holds an
important position in criminal investigations and proceedings. The
polymerase chain reaction (PCR) is often utilized to amplify polymorphic
regions of DNA which are subsequently typed to produce distinct
genotypes. The sensitivity of PCR-based techniques provides a major
advantage over other DNA or conventional serological typing systems.
Samples containing quantities of DNA in the picogram range are often
typed. However, the unprecedented sensitivity of PCR is often cited as a
criticism. One concern is that the interpretation of PCR typing can be
affected by DNA contaminants from foreign sources. In this report, the
level of DNA contamination in New York City Medical Examiner facilities
and its potential affects on HLA-DQA1 typing were assessed. Two related
studies conducted over a five week period measured and typed HLA-DQA1
from accumulated DNA on autopsy room and Forensic DNA Laboratory
structures. The potential for DNA contamination from airborne sources
was also evaluated in the autopsy suites. This study demonstrated the
presence of small amounts of DNA on structural surfaces, but little
evidence of airborne DNA contamination.
10.Riley,DE Compromised optimal hybridization temperatures may explain
PM plus DQA1 anomalies Biochem Mol Biol Int1998 Jan.44(1):59-68:The
multiplex DNA typing test kit, PM plus DQA1 relies on 36 different
oligonucleotides that all must hybridize correctly for accurate DNA
typing. The G + C contents of the probe oligonucleotides range from 47%
to 83%. Since G + C content directly affects optimal hybridization
temperatures, a wide range of optimal hybridization temperatures is
expected. This was verified by two empirically tested formulas. The
recommended hybridization temperature is 55 degrees C. The control probe
has an estimated optimal temperature of 75 degrees C indicating low
stringency of the control. The B allele probe of the GC locus has an
estimated optimal hybridization temperature of 70 degrees C. These
discrepancies are unexplained by GC clusters or sequence palindromes.
The departures from temperature optima may be responsible for observed
PM plus DQA1 artifacts and for previously observed, low tolerance of
this system to temperature variation. Since optimal temperature and
cation concentration are related variables, careful attention to these
is recommended as are positive controls demonstrating amplification of
all alleles monitored by PM plus DQA1.
11. Crouse CA, Vincek V, Caraballo BK Analysis and interpretation of the
HLA DQ alpha "1.1 weak-signal" observed during the PCR-based
typing method J Forensic Sci 1994 Jan; 39(1):41-51 The Perkin-Elmer (PE)
AmpliType DQ alpha Forensic Kit is currently available for amplification
and typing of a polymorphic region in the Human Leukocyte Antigen (HLA)
DQ alpha DNA sequence. Following amplification of the DQ alpha region
with the PE kit, typing strips are processed. These strips contain
immobilized DNA probes designed to distinguish six possible HLA DQ alpha
alleles. It has been observed in this laboratory and others that in a
single source DNA sample, it is possible to detect a weak signal on the
1.1 specific allele dot when the samples' genotype clearly does not
contain the 1.1 allele. It has been suggested that a potential source of
this weak-signal is the non-specific amplification of a HLA DX alpha
gene sequence. To demonstrate the relationship of the DX alpha gene to
the 1.1 nonspecific signal, we designed biotinylated DX alpha PCR
primers specific for a 178 bp region in which the amplified product
spans the homologous DQ alpha region encompassing the DNA probes present
on the typing strips. DX alpha DNA sequences from various DQ alpha
genotypes were amplified and hybridized to DQ alpha typing strips. We
have demonstrated that DX alpha PCR products do not always hybridize to
the 1.1 probe on the typing strips. Sequence analysis of DX alpha PCR
products show that this region is polymorphic which may explain why the
occurrence of the "1.1 weak-signal" is unpredictable. We have
further analyzed the effect of DNA template concentration for the DQ
alpha amplification protocol and have shown that regulation of PCR input
DNA optimizes the amplification and typing protocols for HLA DQ alpha
alleles and minimizes the potential observation of the "1.1
weak-signal
12. Grow M, Phillips V, Reynolds R Post-amplification primer extension
of heat-denatured AmpliType PCR products: effects on typing results J
Forensic Sci 1996 May;41(3):497-502:Alleles of the HLA, DQA1, LDLR, GYPA,
HGBB, D7S8 and GC loci, which are amplified using the AmpliType(R) PM
PCR Reaction Mix and Primer Set, can be detected using sequence-specific
oligonucleotide probes immobilized on a nylon membrane strip. Using
reagents supplied in AmpliType PCR Amplification and Typing Kits,
patterns of blue dots corresponding to particular alleles are visualized
on the DNA probe strips. Frequently, the correct interpretation of
typing results is dependent not only on the presence of probe signals
but also on their relative intensities. The relative probe signal
intensities obtained from an undegraded DNA sample extracted from a
single individual will be different from those obtained from degraded
DNA and from samples containing DNA from more than one source. Because
probe signal intensity is an essential consideration for interpretation,
factors that can influence it need to be identified. Clearly, the time
and temperature of the assay steps and the salt concentration in the
typing solutions can affect probe signal intensity. Also, if
heat-denatured PCR products are allowed to cool for several minutes, the
strands will reanneal and become unavailable for binding to the probes
immobilized on the strips. However, the selective loss of GC B and HLA
DQA1 4.1 probe signals observed after shorter cooling times cannot be
explained by these factors. We demonstrate that following heat
denaturation of PM PCR products there is sufficient residual Taq DNA
polymerase activity to extend primers as the solution cools and that
this primer extension occurs at a more rapid rate than PCR product
reannealing. Primer extension across probe binding sites will prevent
hybridization of the PCR product to complementary probes on the strip.
The extent of signal reduction is dependent on the position of the probe
binding site relative to the 3' ends of the primers and on the strand to
which the probe is complementary. We recommend a simple modification to
the AmpliType typing protocol to ensure all probe binding sites will be
available for hybridization to PM and HLA DQA1 DNA probe strips.
13. Walsh PS, Erlich HA, Higuchi R Preferential PCR amplification of
alleles: mechanisms and solutions PCR Methods Appl 1992 May;
1(4):241-250 :The preferential PCR amplification of one allele relative
to another in a heterozygous sample could result in an incorrect or
ambiguous genetic typing of that sample. There are several mechanisms
that could potentially lead to such preferential PCR amplification.
First, preferential amplification can result from significant GC%
differences between alleles if the conditions of the reaction (denaturation
temperature (Tden), duration at the Tden' salt and co-solvent
concentrations, etc.) allow the denaturation of one allele but not the
other (differential denaturation). For example, the DQa1.1, -1.2, and
-1.3 alleles of the HLA-DQa locus do not amplify at a Tden < 89
degrees C; these same conditions still allow amplification of the DQa2,
-3, and -4 alleles. However, no differences in amplification efficiency
were found between the different HLA-DQa alleles when the Tden was set
at the recommended Tden of 94 degrees C, even after as many as 102
cycles of amplification. Second, for PCR-based genetic typing systems in
which the PCR products from different alleles differ in length,
preferential amplification of the shorter allelic product can occur.
Experiments in which the variable number tandem repeat (VNTR) marker
D17S5 (YNZ22) was amplified under various conditions suggest that the
smaller allelic products are amplified preferentially when Taq
polymerase is limiting. Preferential amplification of VNTR alleles can
also occur if the target DNA is sufficiently degraded. Third, when the
initial number of genomes sampled is very small, stochastic fluctuation
in the number of copies of each allele can result in what appears to be
preferential amplification. Finally, less efficient priming of DNA
synthesis of one allele versus another can occur because of mismatches
between the primer and the specific allelic template, resulting in
preferential amplification of the other allele. General strategies to
avoid preferential amplification are discussed
14.Pai CY, Chou SL, Tang TK, Wei YH, Yang CH Prevention of false results
from preferential PCR amplification of VNTR alleles. J Formos Med Assoc
1996 Jan;95(1):69-72:In forensic DNA typing, evidential samples
generally involve limited amounts of DNA and so should be carefully
utilized. Although polymerase chain reaction (PCR) of variable number of
tandem repeats (VNTR) alleles is the prevailing method for forensic
identification, the fidelity of amplification of heterozygous VNTR
alleles with large disparities in length needs to be carefully examined.
Reports in the literature and our own observations have demonstrated
that PCR artifacts, bogus alleles and allelic drop-out of VNTRs, are
related to the amount of genomic DNA, the number of amplification cycles
and the length of alleles amplified. Two small (< 1 kb) hypervariable
VNTRs (Apo B and HVR-Ig) markers used for forensic identification were
chosen to study these relationships. The results revealed that PCR
amplification for the heterozygous VNTR alleles with wide disparity in
length (> 400 bp) easily produced the allelic drop-out problem and
therefore, led to the false results; and the allelic fragment of PCR
products was preferentially lost after only 2 cycles of
overamplification. We also further established the relationship between
the optimal number of amplification cycles and the amount of genomic DNA
in the reaction mixture. In our routine forensic screening this
relationship has been successfully applied to determine the optimal
number of amplification cycles and to avoid the allelic drop-out problem
and achieve fidelity of PCR-VNTR amplification. It has also been used to
investigate forensic casework.
15. Lee HC, Ladd C, Scherczinger CA, Bourke MT Forensic applications of
DNA typing: part 2: collection and preservation of DNA evidence. Am J
Forensic Med Pathol 1998 Mar;19(1):10-18 : The initial stages of
physical evidence examination are pivotal to the successful resolution
of criminal investigations. Recent cases clearly reinforce the notion
that methods of evidence collection and preservation will continue to be
rigorously scrutinized and challenged in court. This article reviews
forensic applications of DNA typing, focusing on the collection and
preservation of biological evidence. Topics addressed include physical
evidence collection at the crime scene, the forensic laboratory, and the
autopsy room. Specific concerns pertaining to different sources of DNA
evidence are discussed, as are special collection methods associated
with various substrates on which the evidence is deposited.
16. McCormack JM, Sherman ML, Maurer DH Quality control for DNA
contamination in laboratories using PCR-based class II HLA typing
methods. Hum Immunol 1997 Apr 15;54(1):82-88 : Quality control for DNA
contamination in laboratories using PCR-based class II HLA typing
methods. control (QC) in laboratories performing molecular
histocompatibility class II typing often includes a polymerase chain
reaction (PCR) approach for monitoring DNA contamination. An
oligonucleotide primer set was designed, (RBQBf/RBQBr), which is
specific for nonpolymorphic regions of the DR-B, DQ-B, and DP-B
consensus sequences with an expected PCR product size of 81 bp. RBQBf/RBQBr
detected genomic DNA from reference cell lines LWAGS and BM21 (50 to 100
picograms) as well as DR-B, DP-B, and DQ-B amplicon (1 copy).
Additionally, RBQBf/RBQBr detected SSP products from routine DR-B and DQ-B
typings. Validation studies employing controlled DNA contamination of
laboratory surfaces revealed that increasing amounts of wipe test
samples (5% to 20% v/v) were inhibitory to the wipe test PCR, whereas
lower amounts (1% to 2%), or alternatively, a diluted wipe test sample,
increased the sensitivity of the test and optimized the results.
Collectively, this study describes a primer set, RBQBf/RBQBr, which
detects both genomic DNA and DR-B, DQ-B, or DP-B amplicon and
furthermore illustrates the necessity of routine testing for potential
inhibitory factors that may be introduced into the wipe test PCR.
16. Meldgaard M, Morling N Detection and quantitative characterization
of artificial extra peaks following polymerase chain reaction
amplification of 14 short tandem repeat systems used in forensic
investigations Electrophoresis 1997 Oct;18(11):1928-1935 :Detection on
automated DNA sequencers of polymerase chain reaction (PCR) products of
tetra- and penta-nucleotide short tandem repeat (STR) loci frequently
reveals one or more extra peaks along with the true, major allele peak.
The most frequent extra peak pattern is a single smaller peak which is
one repeat unit shorter than the true allele peak. The existence of such
artificial peaks is of special importance when the methods are used for
forensic investigations because the artificial extra peaks may simulate
true alleles when samples containing mixtures of DNA from different
individuals are analyzed. We have investigated the relative levels of
formation of extra peaks in 14 STR marker systems. We found that not
only the parameters of the PCR but also factors determining the
stringency during the post-PCR and pre-electrophoresis handling of
samples were of importance for the formation of extra peaks. In our
hands, the amounts of extra peaks were reduced (i) if the samples were
effectively denatured immediately before loading, (ii) if they contained
substantial amounts of formamide (i.e. > or = 50%), and (iii) if the
temperature of the electrophoresis gel was above a certain level (i.e.
> or = 43 degrees C). The results suggest that extra peaks may in
part be due to re-annealing of the PCR product under suboptimal
conditions.
17. Schmitt C, Schmutzler A, Prinz M, Staak M High sensitive DNA typing
approaches for the analysis of forensic evidence: comparison of nested
variable number of tandem repeats (VNTR) amplification and a short
tandem repeats (STR) polymorphismForensic Sci Int 1994 Jun
3;66(2)129-141: The approach of using nested primers for the APO B
variable number of tandem repeats (VNTR) increases the sensitivity of
the polymerase chain reaction (PCR) to single cell level. Different
experiments and a comparison to the short tandem repeats (STR) system
VWA were carried out, to determine the applicability of this method to
forensic samples. Nested amplification of the Apo B VNTR was affected by
a strong tendency towards preferential amplification of the shorter
alleles. This phenomenon was observed for DNA quantities as low as 100
pg and impaired, depending on the allele length, the results for mixed
samples. As expected, VWA polymorphism showed less preferential
amplification. The high sensitivity of both PCR systems is accompanied
by an increased susceptibility to contamination. Using artificially
contaminated bloodstains, the bloodstain genotype, the contamination or
both genotypes could be found on one piece of evidence. Here a single
analysis can lead to an incorrect result. Therefore a strategy for
obtaining reliable results should consist of multiple stain extractions
and the amplification of different stepped dilutions of the DNA
solution.
18.Wu H, Zhang S, Xiao C Shadow bands in PCR amplification of
trinucleotide repeats and their elimination Chung Hua I Hsueh I Chuan
Hsueh Tsa Chih 1998 Feb 10;15(1):42-45:OBJECTIVE: To explore a method
for eliminating the shadow bands in PCR amplification of trinucleotide
repeats. METHODS: CTG trinucleotide repeat sequence in the
3'-untranslated region of the human myotonin protein kinase gene was
used as templates DNA in PCR amplification, and the effects of Taq DNA
polymerase alone and its mixture with Pwo DNA polymerase on occurrence
of the shadow bands' were compared. The PCR products containing(CTG)5,
both from Taq DNA polymerase and from Taq+Pwo polymerase amplification,
were cloned into pBluecript KS and sequenced on ALF express TM DNA
sequencer respectively. RESULTS: The PCR of DNA sequence containing CTG
repeats frequently produced a main band (usually darker) and a shadow
band (lighter) that differed from the main band by 3 base pairs when Taq
DNA polymerase was used alone. However, when the mixture of Taq DNA
polymerase and Pwo DNA polymerase was used, the shadow bands usually
disappeared. In the 4 clones containing PCR products amplificated by Taq
DNA polymerase, one clone contained only 4 CTG repeats. However, this
kind of decrease of repeat copy was not observed in the 5 clones
containing PCR products amplified by using mixture of Taq+Pwo DNA
polymerase. CONCLUSION: The results of this experiment demonstrate that
shadow bands occur during the PCR. Pwo DNA polymerase in known to
possess not only DNA polymerase activity, but also 3'-5' exonuclease
activity, or proofreading ability. The mixture of Taq DNA polymerase and
Pwo DNA polymerase has higher processivity than Taq DNA polymerase
itself. This explain the effect on eliminating or reducing the
occurrence of shadow bands in polymerase chain reaction
19. Schmerer WM, Hummel S, Herrmann B Reproducibility of aDNA typing Anz
1997 Jun;55(2):199-206 : The reproducibility of short tandem repeat (STR)
amplification of aDNA extracts is limited by means of formation of
artifacts during PCR. One of these artifacts, the so called
allelic-dropout (failure of the amplification of alleles), may result in
false-homozygote typing of a sample, if genotyping is based on the
analysis of a single amplification product. 30 tooth- and bone samples
collected from 17 individuals of three burial sites were investigated in
a blind test. Using the polymerase chain reaction (PCR) the two
independent segregating STR-loci HUMVWA and HUMTH01 were amplified.
Corresponding to a mathematical estimation, multiple amplifications of
each sample and genelocus were carried out. The results of this
genotyping were compared intraindividually.
20.Gallo JC, Thomas E, Novick GE, Herrera RJ Effects of subpopulation
structure on probability calculations of DNA profiles from forensic PCR
analysis. Genetica 1997;101(1):1-12: DNA typing for forensic
identification is a two-step process. The first step involves
determining the profiles of samples collected at the crime scene and
comparing them with the profiles obtained from suspects and the victims.
In the case of a match that includes the suspect as the potential source
of the material collected at the crime scene, the last step in the
process is to answer the question, what is the likelihood that someone
in addition to the suspect could match the profile of the sample
studied? This likelihood is calculated by determining the frequency of
the suspect's profile in the relevant population databases. The design
of forensic databases and the criteria for comparison has been addressed
by the NRC report of 1996 (National Research Council, 1996). However,
the fact that geographical proximity, migrational patterns, and even
cultural and social practices have effects on subpopulation structure
establishes the grounds for further study into its effects on the
calculation of probability of occurrence values. The issue becomes more
relevant in the case of discrete polymorphic markers that show higher
probability of occurrence in the reference populations, where several
orders of magnitude difference between the databases may have an impact
on the jury. In this study, we calculated G values for all possible
pairwise comparisons of allelic frequencies in the different databases
from the races or subpopulations examined. In addition, we analyzed a
set of 24 unrelated Caucasian, 37 unrelated African-American, and 96
unrelated Sioux/Chippewa individuals for seven polymorphic loci (DQA1,
LDLR, GYPA, HBGG, D7S8, GC, and D1S80). All three sets of individuals
where sampled from Minnesota. The probability of occurrence for all
seven loci were calculated with respect to nine different databases:
Caucasian, Arabic, Korean, Sioux/Chippewa, Navajo, Pueblo, African
American, Southeastern Hispanic, and Southwestern Hispanic. Analysis of
the results demonstrated marked differences in the probabilities of
occurrence when individuals were compared to the different populations
and subpopulation databases. The possible genetic and forensic
consequences of subpopulation structu re on probability calculations are
discussed.
21.Budowle B, Koons BW, Moretti TR Forensic Science Research and
Training Center, FBI Laboratory, FBI Academy, J Forensic Sci 1998
May;43(3):657-660 Subtyping of the HLA-DQA1 locus and independence
testing with PM and STR/VNTR loci.. Allele and genotype frequencies for
six loci (HLA-DQA1 and PM loci) were determined in African Americans,
United States Caucasians, and Southwestern Hispanics. The data include
allele frequencies of the HLA-DQA1 4 subtypes. The HLA-DQA1 4 allele
subtyping affords greater power of discrimination in African Americans
and Southwestern Hispanics than in Caucasians, due to the relatively
lower 4.2/4.3 allele frequency in Caucasians. Based on the exact test,
all loci, except the GYPA locus in the African American sample (p =
0.011), meet Hardy-Weinberg expectations. There were two examples of
significant departures from expectations of independence between alleles
of the HLA-DQA1 and PM loci (HBGG/Gc in African Americans, p = 0.30;
LDLR/DQA1 in Caucasians, p = 0.023). The HLA-DQA1 and PM loci also were
tested for associations with three STR loci and the DIS80 locus. There
were four examples of significant departures from expectations of
independence (TPOX/D7S8 and THO1/HBGG in African Americans, p = 0.035
and 0.028, respectively; THO1/LDLR in Caucasians, p = 0.028; and GYPA/D1S80
in Hispanics, p = 0.046). The HLA-DQA1 and PM allele frequency data were
compared with previously reported data on other sample populations of
the same population categories from our laboratory; the allele
frequencies at all loci, except the D7S8 locus in Hispanics (p = 0.028),
were statistically similar. The frequency data can be used in forensic
analyses and paternity tests to estimate the frequency of a multiple
locus DNA profile in various general United States populations.
22. Mueller,L Commentary on Budowle et. al. validation and population
studies of the loci LDLR, GYPA. HBGG,D7S8, and GC(PM loci), and HLA-DQA
(1998) 43 J Forensic Sci 446-447: " A major conclusion of the
(1995) paper by Budowle et. al., was that ‘ all loci meet Hardy-
Weinberg expections and there is little evidence for associations of
alleles between the loci.'.... The genotypes of all individuals in the
FBI's four data bases was provided to me by the FBI through discovery in
a criminal case....It appears that Budowle et. Al. altered the original
data in the Caucasian-HBGG data base when performing (statistical
tests)." Retesting showed " a highly significant departure
from the Hardy- Weinberg equilibrum" for the HGBB locus.
23. Kline, MC Amplication of a vWA Allele (1998) 43 J Forensic Sci 250:
"While typing 600 population samples... our
laboratories...encountered an African American sample that typed
differently at the v WA locus depending on the primers used. The two STR
kits initially used were Promega... Powerplex...and Perkin-Elmer-Applied
Biosystem's AmpF1STR Blue... Subsequently this sample was correctly)
amplified with monoplex vWA primers. The PE-ABI monoplex ... primers
were designed to amplify a shorter product than the vWA primers of the
Amp F1STR Blue multiplex syste. Therefore, these primers do not bind to
the same location on the genome.PE-ABI is aware of this problem and they
are actively pursuing an explanation for this allelic dropout by
sequencing the sample.
24.Harashima N, Liu C, Katsuyama Y, Ota M, Fukushima H Department of
Legal Medicine, Shinshu University School of Medicine.Sequence variation
of allele 27 at the D1S80 locus Int J Legal Med 1997;110(1):22-26: In
180 unrelated Japanese individuals 18 examples of allele 27 were
detected at the locus D1S80 (MCT118). On 6% polylacrylamide gels 5 out
of these 18 alleles were found to migrate between allele 26 and allele
27, but closer to allele 27, and thus were labelled variants of allele
27. All 18 examples of allele 27 were sequenced and the results were
compared. Although all had the same number of base pairs (578 bp) the
five variants could be subdivided into three types. V1, V2 and V3. The
variants and the standard allele were composed of the same kinds of
repeat units, but the order of arrangement was different. We
investigated whether it was possible to distinguish the standard allele
27, and the variants V1, V2, and V3 by PCR-RFLP. EcoRII and MspI which
have restriction sites within the repeat units were adopted as
restriction enzymes. The variants could be discriminated from each other
after treatment of the PCR fragments with EcoRII or MspI, followed by
PAG electrophoresis.
3.THE ABSENCE OF A REGULARLY MAINTAINED PROFICIENCY TESTING PROGRAM
WHICH INCLUDES EXTERNAL PROFICIENCY TRIALS PRECLUDES VAIDATION AND
GENERAL ACCEPTANCE OF THE PCR METHODS USED IN THIS CASE
The importance of external blind proficiency testing on samples that
replicate case work is discussed in sections II E and II F in the
context of developing a reliable method to determine the error rate of a
laboratory. External blind trials, however, also serve a critical role
in establishing the reliability of a new method, or validating that a
laboratory can reliably apply a generally accepted method. The NRC I was
unequivocal on this point:
Most important, there is no substitute for rigorous external proficiency
testing via blind trials. Such proficiency testing constitutes
scientific confirmation that a laboratory's implementation of a method
is valid not only in theory, but in practice. No laboratory should let
its results with a new DNA typing method be used in court, unless it has
undergone such proficiency testing via blind trials.
NRC 1 at 55 (emphasis added).
The Crime Lab has not undergone external proficiency testing via truly
blind trials with any of the PCR based methods. Ironically, it was
established in Bokin that the same Lab does employ blind proficiency
testing for the much less demanding narcotics testing program. Given the
controversy about the reliability of forensic PCR methods among the
leading scientists in the field, the NRC's bright line injunction has
particular force: None of the PCR testing should be used in this case
because the Crime lab has not demonstrated its reliability through
external blind proficiency testing, or, for that matter, through a
regularly maintained open proficiency testing program that includes the
specific kits used in this case. In this regard, the federal auditors
documented in December 1998 that
The record keeping system (of the SFPD) does not facilitate ascertaining
when Analysts have been proficiency tested, for what methods, and with
what results. There is no documentation that the Laboratory Manager has
reviewed the proficiency test results or that the Analyst has seen the
results. In the absence of a Quality Assurance Manager, the Laboratory
Manager must document that proficiency tests have been successfully
completed and that the Analysts have seen their results....Ralph
Whitten's most recent proficiency test result is for 1997. He should be
tested at least every 180 days.
(Exhibit B at 12-13)
Also, it must again be pointed out that Judge Dondero specifically found
in Bokin that "(f)amiliarity with the record involving past efforts
at proficiency testing by Keel suggests lapses in meeting
deadlines." Exhibit A at 14
The NRC I warned that forensic PCR testing "poses even more serious
issues of proficiency, control, and technology transfer than RFLP
typing." NRC Report, at 70. It decried reliance upon the
commercially distributed PCR "kits" that are being used in
this case and observed that "[i]nformation on the extent of the
contamination problem in PCR analysis and the differential amplification
of mixed samples needs to be further developed and published." Id.
Basic methodological issues in the forensic PCR testing remain
unexplored or in controversy. Despite the admonitions of the NRC and
leading scientists, forensic laboratories have not responded to the fact
that techniques for the collection, preservation, and handling of crime
scene samples is an essential part of forensic PCR testing methodology
because of the special problem of sample contamination. Although such
organizations as the American Society for Histocompatibility and
Immunogenetics has, for PCR techniques, detailed "Standards for
Histocompatibility Testing," there is no comparable set of
guidelines for forensic PCR testing. In this envirornment of an
unregulated industry operating without objective standards, a regularly
maintained proficiency testing program is one of the few safeguards
available to ensure correct scientific results. The absence of such a
program in this case precludes a finding of general acceptance of the
Crime Lab's methods and procedures.
B. THE DQ ALPHA+PM TEST KIT USED BY THE SF CRIME LAB IS NOT GENERALLY
ACCEPTED IN THE SCIENTIFIC COMMUNITY
As discussed above, neither Morganti nor Wright consider the general
acceptance of the DQ Alpha+PM test used in this case. This multiplex kit
was not in use at the time of Morganti or Wright. This kit is a
substantial scientific departure from the kits used in Morganti and
Wright. TWGDAM 4.4.1.6 specifically requires as part of the
developmental validation process that " (w)here more than one locus
is amplified in one sample mixture, the effects of such amplification on
each system (alleles) must be addressed and documented." The
defense submits that the DQ AlphaPM test is not generally accepted as
reliable because of a multitude of problems in the following areas of
concern: the quality and specificity of amplification,quantification
problems,inhibition, contamination, false positives,and the subjective
nature of determining genotypes when dealing with mixed samples.
a. Quality and Specificity
Amplification conditions can affect the quality and specificity of the
process. Critical parameters include the temperatures and times period
of the amplification and the number of cycles of amplification; the
composition of the amplification mixture from primer to the chemical
reagents use; and the amount and nature (single stranded or double
stranded) of the DNA from the sample. NRC 1, at 63-64. The thermocycler
(the machine used in the amplification process to control the
temperature of the reaction) can effect the results based on simple
variations in temperature between manufacturers, or from sample well to
sample well, or simply from failure to adequate calibrate and monitor
the accuracy of the thermocycler. Id. at 64. Additionally, for each
assay precise amplification conditions must be established to ensure
reliability and accuracy. NRC 1, at 64.
b.Quality and Quantity
The quality and quantity of the amplified DNA is also a major concern.
Under certain conditions, it is known that DNA polymerase, the enzyme
used to amplify DNA, may "misincorporate" in a random or a
specific fashion. Id. Misincorporation means that during the process of
amplification, incorrect bases are added to the DNA being amplified
resulting in a DNA that has a different sequence than the starting DNA.
Therefore, when the amplified DNA is typed based on its sequence such as
with the PM and DQ tests, there will be errors in the genotype due to
misincorporation.
When the DNA to be amplified contains tandem repeated units, the DNA
polymerase may slip during the amplification resulting in additions or
deletions of the repeat, which creates a heterogeneous collection of
amplified DNA, making the interpretation difficult. Id. Therefore, each
new assay, or each new gene/marker, must be studied for DNA polymerase
misincorporation or slipping prior to its use in the forensic setting.
Amplification may also be affected by the efficiency of the
amplification of different alleles of a marker. Id. For instance, a
sample may contain a 50:50 mixture of two alleles but yield a 90:10
ratio of amplified DNA. This "differential amplification" may
be the result of a number of mechanisms, including the length of the
sample DNA (larger fragments amplify less efficiently); the difference
in G-C nucleotide content of a sequence which affects the denaturation
efficiency (more G and C nucleotides in a sequence results in lower
denaturation); and the temperature of the reaction. Id. In order to
ensure that differential amplification does not occur for a given
allele, there must be safeguards in place for each test and in every
laboratory.
Differential amplification is a particular problem where there is a
mixed sample. For example, where the DNA is typed and four alleles are
seen, with alleles 1 and 2 being stronger than 3 and 4. The conclusion
that the alleles 1 and 2 come from one source while 3 and 4 from a
second could only be justified if it has been demonstrated that the
amplification process for that gene yielded amplified DNA in direct
proportion to the starting DNA. Id. at 65. If there was differential
amplification, then no conclusion as to which allele came from the same
source would not be justified.
Finally, differential amplification is seen where the primer amplifies
DNA that does not come from the target gene but from some other random
sequence of DNA. Id. This occurs when the number of amplification cycles
is long, therefore, it is necessary to determine the limit of cycles for
each allele.
c.Inhibition
Forensic samples may contain factors that inhibit the amplification.
This happens by the factor binding to the DNA or by inhibiting the DNA
polymerase. Id. at 65. This is seen in DNA from old blood stains and
from specific fabrics or material, for example, blue denim. Id.; Del
Rio, et al, PCR-Based Human Leukocyte Antigen (HLA) DQ Typing of Blood
Stained Light and Dark Blue Denim Fabric, (1996) 40(3) 490. "Each
PCR system should be thoroughly characterized on a range of simulated
and known forensic samples, to document any effect on reliability. NRC
1, at 65.
d.Contamination
As discussed at length above, contamination of samples can limit the
reliability of any PCR amplification. Contamination of samples equally
affects the reliability of the new PM+DQA1 kit used in this case.
e. False Positives
As indicated above, one recent article regarding the combined DQ Alpha
+PM kit found that attempts at typing samples using this multiplex kit
resulted in failed hybridization and allelic dropout. Donald E. Riley,
Compromised Optimal Hybridization Temperatures May Explain PM Plus DAQ1
Anomalies, 44 Biochemistry and Molecular Biology International 59, 63
(1998). Surveying prior studies, Dr. Riley states that "three
independent studies reported false typings with either DQ Alpha and
Polymarker. In two of these studies, the false typings were confirmed by
other techniques such as DNA sequence data or by comparisions with
nonmultiplex systems."
f. The PM+DQA1 Test Kit is Not Reliable and Therefore Not Generally
Accepted Because of the Subjective Nature of Determining Genotype with
Mixed Samples
No matter the procedure used to detect amplified DNA, there are risks
that the resulting genotype will be misinterpreted. Id. at 68.
Reverse dot blotting is used in the test for DQ and PM. The absence of a
dot theoretically means that the allele is not present. Id. The presence
of a dot generally means that an allele is present. However, the
presence of a intermediate intensity dot, can be the result of a number
of events. There could be contamination from another sample or the
technician, or there could be technical variations in the condition of
the particular test, or the weaker dot shows a true heterozygote. This
type of problem can be resolved by repeating the experiment or comparing
the sample at a number of loci or markers. "Rarely, the origin of
an intermediate hybridization signal might remain unresolved; if the
type of the sample is still questionable, the data should be
discarded." Id.
An illustration of the problems encountered with determining a genotype
from reverse dot blot strips is seen in a study in which field tests
were conducted by seven laboratories using the PM+DQA1 kit. Fildes,
Nicola, and Reynolds, Rebecca, Consistency and Reproducibility of
AmpliType® PM Results Between Seven Laboratories: Field Trial Results,
Journal of Forensic Sciences (March 1995) 40:279. Of the 287
"DNA-containing samples" tested for the PM markers, there were
7 samples in which no results were reported or the results were "uninterpretable."
Of the 224 samples tested for DQ , 14 samples were reported as either no
results or uninterpretable results. Id. at 281. The "no result
reported" are those samples in which there may have been some
typing dots visible, but no control dot was visible. Id. The "uninterpretable
result reported" samples were those in which the
"interpretation of the strip was prevented by extraneous dots on
either DNA probe strip or by imbalanced dots on the PM DNA probe strip
that were greater in intensity than the ‘S' and/or ‘C' dots."
Id. at 282. An "extraneous dot" is one that indicates the
presence of more than two alleles, or a mixed sample. In general, these
results reflected poor separation of sperm and epithelial cells in
"postcoital swab" extracts, however, there were at least eight
strips that exhibited contamination from external sources. Id. at 282,
284.
The laboratories in the study were requested to repeat the tests for the
simulated sexual assault samples. Id. Two of the six laboratories that
repeated the tests did not reproduce the same results. Id. One
laboratory obtained 100% reproducibility in the PM test, but only 82% in
the DQ test. The second laboratory achieved on 91% reproducibility in
the PM test and 82% reproducibility in the DQ test. Id.
The PM test is designed so that the dot intensities are balanced, so
that if there is a mixture, then the darker dots are paired and the
lighter dots are paired. The PM test is sensitive to variations in the
alleles of certain markers and to assay conditions. The GYPA locus has
variants that affect the intensity of the dots, making it difficult to
type the strips due to the color imbalance. Id. at 284. If the
conditions of the assay fall outside the parameters of the method, such
as differences in the water bath temperature, salt concentrations of
solution or times of incubation, the balance of the dot intensities may
be affected. Id. The imbalance caused by the conditions of the test also
affects the balance of the "S" control dot, making it
difficult to interpret the results of a mixed sample.
In 1996, a study was published that outlined the problem of reduced
signal intensity at the GC B allele and the DQ 4.1 allele that could not
be attributed solely to renaturation of the DNA prior to hybridization
to the strip. Grow et al., Post-Amplification Primer Extension of
Heat-Denatured AmpliType® PCR Products: Effects on Typing Results,
Journal of Forensic Sciences (May 1996) 41:497. The study concluded that
after the amplification was completed, these two alleles continued to be
extended resulting in a decreased signal on the strip, and that the heat
treatment recommended was not sufficient to prevent the decrease. Id. at
500.
5. THE D1S80 TEST KIT USED BY THE CRIME LAB IS NOT GENERALLY ACCEPTED AS
RELIABLE IN THE SCIENTIFIC COMMUNITY.
The D1S80 testing in this case was performed using a AmpliFLP D1S80
PCR Amplification and Typing Kit ("D1S80 kit") marketed by
Perkin-Elmer. These kits first appeared in 1993.
The kit was specifically designed to attempt to type DNA for forensic
identification purposes. The manufacturer specifies that the kit is for
forensic or research use only and is not for use in diagnostic
procedures. See Perkin-Elmer AmpliFLP D1S80 Allelic Ladder Kit Insert.
Thus, the manufacturer recognizes that the kit is unsuitable for medical
and clinical uses, which require certification, which D1S80 has not
achieved. It can only be distributed in the unregulated, non-certified
crime-lab community. This produces the paradoxical result that
"clinical laboratories must meet higher standards to be allowed to
diagnose strep throat than forensic laboratories must meet to put a
defendant on death row." Eric Lander, DNA Fingerprinting on Trial,
399 Nature 501, 505.
As discussed above, California appellate cases have no considered or
approved the D1S80 procedure. Many labs do not even attempt to test
crime scene evidence with the D1S80 kit.
a. D1S80 was Commercially Developed and had No Independent Testing or
Approval Before Being Marketed and Used
The use of a commercially distributed D1S80 kit does not mean that
the test is valid or reliable for use for crime scene evidence. There is
no regulatory monitoring or control of this kit. It is not officially or
independently tested or approved before being introduced commercially
for use in crime laboratories. The NRC has recommended that kits be
tested and approved by and independent government agency before they are
put to use by the crime labs. NRC 1 at 69; see also Randolph H. JonaKait,
Forensic Science: The Need for Regulation, 4 Harvard J. of Law and
Technology 109-140 (1991); Office of Technology Assessment, Genetic
Witness (1990).
b. D1S80 Testing of Mixed Crime Scene Evidence Samples from Unknown
Sources is not Generally Accepted as Valid and Reliable
Applying DNA testing methods to crime scene evidence is much more
difficult than simply using the methods developed for research purposes
with known and undegraded or uncontaminated samples. The techniques are
more demanding with crime scene evidence than with other applications of
the same techniques. "It is generally accepted that applying DNA
tests to forensic samples, especially criminal evidence, potentially
presents more difficulties than analyzing samples in basic research or
clinical diagnosis. Samples from crime scenes are frequently small and
might be of poor quality because of exposure to a spectrum of
environmental onslaughts. Unlike paternity samples, where each sample is
from an identified source, the contributor to evidence from a crime
scene is often unknown." Office of Technology Assessment, supra at
59. See also Lander, supra, at 501 (use of DNA test analysis for medical
diagnosis does not mean the technique is reliable for forensic
identification).
The D1S80 method, as applied to testing crime scene evidence, involves
additional critical steps that do not arise in other applications of the
technology of amplifying non-crime scene samples. NRC 1 at 77. For
example, the NRC 1 report noted that the scientific community was
uniformly alarmed about the special dangers of contamination in crime
scene evidence. NRC 1 at 65-66.
As indicated above, to be accepted under Kelly, the D1S80 method must be
accepted as reliable as it is applied in the specific forensic
application. For the D1S80 typing kit used in this case, there have been
published validation studies for the D1S80 kit. Nevertheless, the
experience with this kit has shown that there are serious reliability
problems which prevent the kit from being generally accepted in the
community of molecular biologists.
1. the microbial contamination problem
The most serious problem to emerge is the inability of the kit to
distinguish between human DNA and certain forms of microbial DNA
commonly found in crime scene samples. All experts for both sides in
Bokin agreed that this is a serious concern. On this issue, the People's
Department of Justice witness, Dr. John Tonkyn, testified as follows:
b. You said that the TWGDAM guidelines do not specifically call for
microbial studies?
a. Not for internal validation, no.
1. But for developmental validation it's widely understood in the
forensic community that nonhuman studies, including microbial studies,
are very important when it comes to STR typing?
1. That's correct.
17. As well as D1S80 or any PCR system?
1. Yes.
Q. And the reason for that is it's known generally that microbial DNA
can produce bands that, in some situations, can mimic human bands?
1. I think the reason that that study is included in TWGDAM is that
forensic samples are often contaminated with microbial DNA, and you want
to determine if the system that you're using can produce bands that you
might mistake for those created from human DNA.
(R.T. 5400)
Although the People's own witness thus acknowledged the importance of
the microbial contamination problem, the People in Bokin cited to
pre-1996 D1S80 validation studies that have generally ignored or failed
to comprehensively consider this issue. The People did acknowledge a
1996 study, Fernandez-Rodriguez, et al., Microbial DNA Challenge Studies
of PCR-Based Systems Used in Forensic Genetics, Advances in Forensic
Haemogenetics, 16th Congress of the Interational Society for Forensic
Haemogenetics 177 (1996), which was the first to document the microbial
contamination problem, but the People failed to appreciate the
significance of the study. The study found that certain microbes
commonly found in crime scene samples do in fact mimic human DNA. A 1997
study by the same authors repeated and expanded the first study to
include additional microbes.(Defense Exhibit TT) The People in Bokin did
not cite the more recent study, but they did provide a faxed letter by
the authors of the study written after the first study, but before the
second, indicating that Dr. Bruce Budowle of the FBI, under unknown
circumstances, prevailed upon the authors, who are themselves law
enforcement personnel engaged in routine D1S80 forensic casework, to
opine that their first study should not preclude forensic use of D1S80
testing methods under certain circumstances, including quantification of
human and total DNA, and " double typing by native and denaturing
PAGE." The more recent study found additional microbes that mimiced
human DNA and also suggested that the problem could be solved by double
typing by native and denaturing PAGE.
The short answer to any reliance on the faxed letter is that it is
irrelevant because Allen Keel does not in fact utilize the suggested
remedial procedure. More broadly, as in People v Brown, 40 Cal 3d at
533, the People cannot seriously contend that the hearsay opinion letter
of forensic technicians is sufficient to state the views " of the
relevant community of impartial scientists."(emphasis in original)
Those scientists, including Drs. Reilly and Libby, are of the strong
opinion that the microbial contamination problem precludes the
unmodified D1S80 kit methodology from being generally accepted within
the community of molecular biologists.
The Kelly-Frye admissibility hearings conducted to date on the D1S80 kit
have focussed heavily on the microbial contamination problem. In
Riverside County, following a lengthy hearing, Superior Court Judge
Bailey declared that results of the D1S80 test were inadmissible because
" (t)he Court finds that there is a genuine controversy, difference
of opinion and lack of consensus in the relevant scientific community as
to the general acceptance and effectiveness of the PCR based D1S80
testing procedure." People v. Scott, Ruling of Jan. 23, 1997,
submitted as Exhibit C in this hearing, at p. 6. Central to the Court's
ruling was the finding that the microbial contamination problem
identified in the Fernandez- Rodriquez study " has not been
adequately provided for in the D1S80 kit now distributed by Roche
Molecular Systems and appears not to be adequately addressed by others
doing the D1S80 testing procedure." (Id.)
In San Francisco, the D1S80 test was addressed in Bokin, but not in
McClanahan. Judge Dondero opted to treat the microbial contamination
problem as a prong three issue rather than a prong one issue as was done
in Scott. He wrote: " The fact the (Spainish) researchers found
continued evidence of microbial DNA which mimicked himan DNA does not
cause this Court toexclude such evidence, especially when the problem
can be controlled by certain measures int he particular laboratory. Any
forensic witness subpeoned by the prosecution would be expected to
consider and deal with this observation by the Madrid group."
Ruling of May 6, 1999, submitted as exhibit A in this hearing.
Therefore, whether the issue is treated under prong 1 or prong 3 of
Kelly- Frye, it is clear that the microbial contamination problem
requires some showing by the People that the problem has been resolved
by adopting the procedures suggested in the Spainish studies or some
other equally effective procedure. No court yet, however, has held that
the problem can simply be ignored on the say so of a particular lab
technician who believes, contrary to the weight of scientific opinion,
that the problem does not merit solving.
2. Other Reliability Problems
Moreover, there are also numerous other reliability problems with the
D1S80 typing kit which precludes its general acceptance. The peer
reviewed scientific articles marked as defense exhibits in Bokin
highlight a few of those problems, including that " many variants
(or interalleles) exist in the D1s80 locus that do not match the allelic
ladder, which makes genotype identification difficult"(Exhibit 57),
that the effectiveness of the method is " often reduced by the
occurrence of PCR artifacts, such as bogus alleles and allelic
dropout" (exhibit 58), that " more recent studies indicate
that it is possible to obtain ambiguous and even incorrect results from
moderately degraded DNA samples" (Exhibit 64), that, in the words
of the kit insert, "[a]t certain levels of DNA degradation, only
the smaller allele of a heterozygous sample may be detected (and) [s]ome
amplified D1s80 alleles in the size range of the ...Allelic Ladder do
not line up with a ladder band" (Exhibit 65), and that "
preferential amplification, anomalous products , and length limitations
may be incountered" (Exhibit 66), and finally, in the words of Dr.
Holt and her colleages, that " [w]ith regard to the D1S80 kit, the
discrimination potentail of this locus can be very informative, yet the
allele size range is rather large,resulting in potential for
preferential amplification." (People's Exhibit 23). These problems
go beyond " issues and system limitations" as the People
euphemistically put it in their brief in Bokin. (Brief, p 9-10) They go
to "the crucial issue:whether for the typing categor(y)...at issue
here, current methodology, employed by qualified techniquians, can
discriminate reliably between testable and untestable samples and
between accurate and inaccurate results." People v. Brown , 40 Cal.
3d at 534 (emphasis in original) The literature and the testimony before
the Court is that there is currently no scientific consensus on this
crucial issue.
c. The D1S80 Test Does Not Contain Basic Scientific Controls
To be scientifically accepted, a test must incorporate controls to
identify internal failures and contamination and otherwise assure that
the test procedures are working properly and producing correct results.
There are a number of types of controls designed to detect different
potential problems. Positive controls show whether the test correctly
produces positive results when it should. Negative controls show whether
the test correctly produces negative or not results when it should.
Substrate controls show whether or not a sample has been contaminated by
its substrate and/or handling.
The scientific method and principles of validity and reliability dictate
that no meaningful conclusions can be drawn from a test unless all
appropriate controls have been included. As the National Research
Council noted, "[f]orensic DNA analysis should be governed by the
highest standards of scientific rigor in analysis and interpretation.
Such high standards are appropriate for two reasons: the probative power
of DNA typing can be so great that it can outweigh all other evidence in
a trial; and the procedures for DNA typing can be so great that it can
outweigh all other evidence in a trial; and the procedures for DNA
typing are complex, and judges and juries cannot properly weigh and
evaluate conclusions based on differing standards of rigor. NRC 1 at 52.
See also Randolph N. Jonakait, Will Blood Tell? Genetic Markers in
Criminal Cases, 31 Emory L.J. 833.
D1S80 lacks all the minimum controls that "should be processed at
the same time as the evidence samples." Technical Working Group for
DNA Analysis Methods, Guidelines for a Quality Assurance Program for DNA
Analysis (1995) Crime Lab Digest §7.5 [hereinafter TWGDAM Guidelines].
Controls required by TWGDAM and not included in D1S80 testing are:
1. Negative controls: reagent blank and amplification blank. TWGDAM
Guidelines, §7.5.1.1;
2. Positive Controls: known human DNA "carried through the…
typing…. TWGDAM Guidelines §7.5.1.2;
3. Substrate Controls, which "should be collected from the
evidence, including unstained areas adjacent to stained areas …and
should be processed at the same time as the evidence samples."
TWGDAM Guidelines §7.5.1.3;
4. The TWGDAM guidelines further dictate that "the sample should be
split for duplicate analysis as early as possible prior to the
amplification" as a further control. TWGDAM Guidelines §7.5.1.4.
d. Research Testing of D1S80 has Revealed Many False Positives
Many false results have been documented in the scientific literature.
See, e.g., Kloosterman, et al., PCR Amplification and Detection of the
Human D1S80 Locus, 105 J. Legal Med. 257, 263 (1993). Among the problems
that may cause false identification are:
1. Differential Amplification:
c. The technician my falsely interpret the results as homozygous;
d. Missing alleles in a mixes sample will result in false typing;
2. Degraded DNA will result in false results by failing to resolve all
alleles through electrophoresis, and/or through the addition of false
bands in the results;
3. Mixed samples will likely result in varying band intensity and a
false interpretation;
4. Certain strains of bacteria produce false bands on the human allelic
measurement ladder;
5. There are additional alleles that have not been identified;
6. Insufficient or excessive quantity of DNA can result in a false
signal, and thus a false result that is undetected by the analyst.
Rebecca Reynolds et al CITE.
e. Preferential Amplification
A known and generally accepted principle of the PCR-amplification
process is that smaller alleles are amplified preferentially over larger
ones. TWGDAM Guidelines require that "the potential for [problems
with] differential [preferential] amplification must be addressed.
TWGDAM Guidelines §4.4.1.5. A serious scientific defect of the D1S80
test is its inability to specifically identify all alleles.
Preferential amplification proses and insurmountable problem when
attempting to interpret a mixture. An analyst cannot determine whether
one of the larger alleles was "out-competed" in the
amplification process. What the analyst sees in interpreting a mixture
is potentially incorrect. Furthermore, the test contains no internal
control or other device to identify when or if preferential
amplification has occurred.
Similarly, there are no controls for "off ladder variants,"
those alleles that do not fall with the core-repeats allelic ladder.
One study that attempted to validate the D1S80 test, reported numerous
problems in achieving reliable results. In a paper released following a
presentation at the 14th Congress of the International Society for
Forensic Haemogenetics, the authors concluded that using the D1S80 test
should only be undertaken and interpreted "with great care,"
and that "more recent studies indicate that it is possible to
obtain ambiguous and even incorrect D1S80 types from moderately degraded
D1S80 samples." Marco A Scarpetti & Stephanie A. Turek,
Validation of the AmpliFLP D1S80 Amplification Kit for Casework in the
Detroit Police Crime Lab, Sixth International Symposium on Human
Identification (Promega Conference) Scottsdale, Arizona, October, 1995,
Poster Abstract.
The authors of this paper also stated that "[a]mbiguous types are
obtained when the smaller of two observed alleles is amplified to a
significantly greater extent than the other allele. This type of result
can arise from amplification of a partially degraded DNA sample or from
amplification of one of a variety of mixed samples." Id. The
authors went on to warn that "[i]ncorrect types are obtained when
the degraded DNA sample can support amplification of products in the
size range of the smaller allele but not in the size range of the larger
allele. As a result, a sample from a heterozygous individual can be
analyzed incorrectly as a homozygous type." Id. The authors
discussed the inadequacy of the D1S80 ladder and detailed the ongoing
development and release of new component ingredients.
6. THE GREEN I AND PROFILER PLUS STR TEST KITS, USED IN CONJUNCTION
WITH THE 310 GENETIC ANALIZER AND AUTOMATED CAPILLARY ELECTRPHORESIS,ARE
NOT GENERALLY ACCEPTED IN THE SCIENTIFIC COMMUNITY
In their pleadings in Bokin and McClanahan, the People provided a list
of 34 "articles" that purport to support their contention that
STR typing of forensic samples is generally accepted in the scientific
community. Although "the People are not aware of any publications
questioning the scientific reliability of PCR-based DNA analyses and
databases" Opp. at 76, several of the articles included in their
list of publications detail many of the problems that arise from
amplification and typing STR loci. See, e.g., T.M. Clayton et al.,
Further Validation of a Quadruplex STR DNA Typing System: A
Collaborative Effort to Identify Victims of a Mass Disaster, For. Sci.
Int'l 17-25 (1995) (finding that only 37% of samples were successfully
typed when four STR loci were amplified simultaneously and a full 44% of
samples yielded no result); K.A. Micka, et al., Validation of Multiplex
Polymorphic STR Amplification Sets Developed for Personal Identification
Applications, 41 J. For. Sci 582-590 (1996) (noting problems in typing
mixtures); P. Gill, et al., Development of Guidelines to Designate
Alleles Using an STR Multiplex System, 89 For. Sci. Int'l 185-97 (1997)
(describing problems with stutter bands and non-specific artifacts).
Most remarkably, not one of the articles listed by the People in their
pleadings verify that the methods used by the Crime Lab in their STR
typing of TPOX, THO1 and CSFIPO loci or the nine Profiler Plus loci are
appropriate for forensic use. Such verification is required by the
National Research Council, and discussed in detail, in section I(D)(5),
supra.
Not one of the articles listed in the People's motion verify that the
tests used and protocols followed by the Crime lab in their typing of
the thirteen STR loci in this case are "valid not only in theory,
but also in practice." Indeed, most of the articles listed are
completely irrelevant to the reliability of the testing conducted by the
Crime Lab. See, e.g., J.L. Weber & P.E. May, Abundant class of Human
DNA Polymorphisms Which Can be Typed Using the Polymerase Chain
Reaction, 44 Am. J. Hum. Gen. 388 (1989); P. Gill, et al., Automated
Short Tandem Repeat (STR) Analysis in Forensic Casework-A Strategy for
the Future 16 Electrophoresis 1543 (1995); Walsh, et. Al Development of
the AmpF1STR Blue PCR Amplification Kit (Perkin-Elmer product
literature).
The People attempt to show general acceptance and reliability by
illustrating that research is conducted in the area of STR typing.
Clearly, abundance of research and general acceptance and reliability
are not the same. Without a doubt, there is , for example, an abundance
of research on the topic of cold fusion, but obviously, there is no
general acceptance of that theory.
Another crucial concept neglected by the People and their witness is the
concept of validation. TWGDAM Guideline 4.1.2 states: "Validation
studies must be conducted by the DNA laboratory or scientific community
prior to the adoption of the procedure by the DNA laboratory."
Contrary to all other witnesses, both NRC reports, and the TWGDAM and
DNA Advisory Board Guidelines, Dr. Cydne Holt began her testimony in
Bokin with the dubious assertion that validation studies "do not
overlap with the idea of general acceptance of STR's necessarily,"
and that she did not need to rely, as she had so extensively done in
McClanahan, on Perkin Elmer's Green I validation study. ( R.T. 46) This
shift of position, of course, was dictated solely by the company's
desire to maintain trade secrets and Judge Dondero's ruling upholding
the priviledge but denying to the People the use of the company's
validation study. But even Dr. Holt had to admit on cross examination
that somebody somewhere has to validate a system before it is used
forensically. (R.T. 96)
Importantly though, Dr. Holt also maintained that the way the validation
process works is that "the manufacturers ...release the
kit...and...by definition the validation studies are performed by the
forensic scientist on the specific commercially available kit,
post-release from the manufacturer." (R.T. 114) Fast forward then
to the testimony of Dr. John Tonkyn, who maintained that the DOJ never
did a full developmental validation study on the Green I kit because
"[t]he developmental validation had already been carried out at
Perkin Elmer, which markets this kit." (R.T. 5340) Like Holt,
Tonkyn testified that "[t]he kit requires 4.1, developmental
validation, before it can be utilized. In addition, each laboratory that
uses this must comply with Section 4.5, internal validation." (R.T.
5456) Significantly, Tonkyn conceded on cross examination that "a
validation study which dealt, for instance, with the Promega CTT system,
using the Promega primers in a silver staining system, would not
validate developmentally or internally the Green I Kit, which uses an
entirely different set of primers, an entirely different software, and
entirely different parameters." (R.T. 5477) The People's own
witness thus completely undermines the People's reliance in its brief in
Bokin on validation studies such as those contained at Tabs 2, 6-7,
9-10, and 20 of the People's post-hearing brief, all of which are
subject to the same criticism leveled by Tonkyn against a damaging
stutter study brought to his attention by the defense.
This leaves only Tab 8, The Perkin Elmer Manual that the Court ruled the
People could not rely upon because of Perkin Elmer's refusal to disclose
the underlying data. When Dr. Tonkyn was specifically asked to cite all
of the validation studies he was aware of concerning the Green I kit,
the only one he could cite was Perkin Elmer's Green I Manual. (R.T. 5477
) The Manual, in turn, cites environmental abuse and nonhuman specimen
studies supposedly conducted by DOJ which Tonkyn swears DOJ never
conducted. (R.T. 5341, 5383,5394-5397) Despite the fact the Manual cites
studies that were never conducted, and that the data underlying the
Manual has never been examined by DOJ or anyone else, Tonkyn testified
that when the DOJ first performed casework in September of 1997
utilizing the Green kit and the 377 Sequencer (DOJ never did casework
using the 310; R.T. 5479) "we relied on the developmental
validation that had gone into that User's Manual to cover Section 4.1 of
TWGDAM." (R.T. 5454)
Another disturbing aspect of the testimony of Dr. Holt, representing
perhaps the best example of her bias and downright propensity to
fabricate testimony, is her testimony in McClanahan and Bokin about the
American Academy of Forensic Sciences abstract entitled TWGDAM
Validation of AmpFLSTR Green I and AmpFLSTR Profiler I by Dr. John
Tonkyn ( People' Exhibit 80.)
In McClanahan, the People introduced the abstract as one basis for Dr.
Holt's broad assertion that "multiple laboratories have done our
experiments over again and have validated the green and all the other
multiplexes that we've developed." (R.T. 831) She specifically
testified that this abstract was "definitely peer reviewed" (R.T.
834) and "I've read it and I was there when he presented
it..." (R.T. 916) She continued: "Dr. Tonkyn is representing
the California Department of Justice in this talk and their validation
found Green I and Profiler to be useful for forensic investigations.
They now use these systems routinely in their laboratory. They performed
experiments to address all the TWGDAM guideline that we've been going
over." (R.T. 915)
In the consolidated cases, Holt was again called upon by the People to
emphasize the importance of Dr. Tonkyn's abstract:
17. And does that particular abstract have any significance in terms of
your opinion that STRs are generally accepted and reliable for human
identification?
1. Sure.
17. Are there any parts in particular without reading the entire
abstract?
1. This scientist validated, as the titles says, according to TWGDAM
guidelines the Green I and the Profiler Kit for forensic casework, both
of those kits contain the STR loci that we have been talking about, CTT,
using the same exact reaction conditions and primer pairs. They address
the TWGDAM guidelines. I won't read them all out unless the court would
like to hear those again. Some of the particular results that they were
presenting are effects of environmental abuse, such as semen and
bloodstains to several temperatures and sunlight, effect of degraded
DNA, and effects of PCR inhibitors, and in all cases, the correct
genotype was obtained. ... And they are using a minimum signal threshold
of 150 like we have been talking about. And go on to generalize:
"that based on the results of our validation study, guidelines were
developed which take into account signal strength, the presence of
stutter bands and peak height rations to assist in the interpretation of
casework and offender databank results and the statistical analysis of
genotype frequency."
q. So would it be fair to say that the California Department of
Justice advertise validated the Green I Kit for forensic use?
a. Yes.
(R.T. 963)
On cross examination, Dr. Holt further gilded the lily by claiming
that the DOJ abstract concerning Green I and Profiler I was not a work
in progress, but rather, and more significantly, "a report of their
validation studies. They had finished those when they presented it....
They validated the system. Their efforts or their experiments--It was
already finished when they presented their work." (R.T. 965) She
even testified that she "remembered Dr. Tonkyn presenting their
guidelines relative to quantitating the stutter and peak height ratios
for interpreting mixtures." (R.T. 988)
Judge Dondero soon learned that most, if not all, of this testimony is
an utter fabrication. Following Dr. Holt's testimony, and in direct
contradiction to it, the Attorney General inadvertently undermined Dr.
Holt's testimony by presented the declaration of DOJ Lab Director Lance
Gima, which declares under penalty of perjury that "I am
specifically aware of the validation studies for the AmpFISTR Green I
and... Profiler I conducted by DOJ analysts....These validation studies
are internal rather than developmental. As such, they are primarily
pertinent to this laboratory's assessment of the effectiveness of our
own procedures and methodology. The studies have not, as yet, been peer
reviewed....Further, with respect to AmpFISTR Profiler I, DOJ cannot
provide a complete validation study since our laboratory did not
complete the study; instead, our laboratory began using...Profiler II
during the development of our multiplex protocols, and it was that data
which was presented at the February 1998 American Academy of Forensic
Sciences proceedings. Profiler II employs different loci than does
Profiler I." (Defense Exhibit 7K)
Then, apparently forgetting or choosing to ignore the testimony of Dr.
Holt, the People presented the rebuttal testimony of Dr. Tonkyn to the
effect that the DOJ Green I study was not in fact a developmental
validation study at all, but only an internal one, and that it was done
not on a capillary electrophoresis platform using a 310 Genetic
Analyzer, but on a gel-based 377 DNA Sequencer using a 1996 pre-release
kit and "in-house procedures" that were significantly modified
from the protocol in the 1997 Green I User's Manual.(R.T. 5340-5341;
5421-5424, 5432, 5453). As an example, Tonkyn stated that the study was
conducted with a stutter cut-off of twenty percent, not the seven/ten
percent cut-offs which are currently used, which means that the study
was eliminating calls in the ten to nineteen percent range which the
current program would call as true alleles. (R.T. 5360) He also admitted
that the study was flawed because of "a mistake that I had made in
interpreting a peak as stutter, when, upon reanalysis, we are positive
that this represents a spillover artifact from one lane of a gel into
the next lane." He further stated that the study had found a number
of atypical results, including stutter peaks in excess of the ten/seven
cut-off, and one stutter peak as high as twenty two percent (which he
now attributed to his carelessness in causing contamination and improper
interpretation), as well as a null allele, imbalanced alleles, and
alleles falling below the expected 150 RFU cut-off. (R.T. 5374,
5387-5392)
There were other problems as well. In the mixture portion of the study,
the samples were "overlapped each other so much that the whole
experiment was sort of a wash. You couldn't really draw any
conclusions." (R.T. 5368) Also, "Perkin Elmer had not
finalized the protocol which included the particular number of cycles of
PCR." (Id). Moreover, there was a major problem with the
quantification of the test samples, as a DOJ analyst "had
quantified or quantitated the samples at a much earlier time point. And
what we have seen when you have DNA stored at very low concentrations,
the concentration will sometimes go off." (R.T. 5399) Along the
same lines, in the population database portion of the study, for some
samples "the actual quantitation of the DNA was done incorrectly or
the DNA degraded during shipping, possibly due to not being frozen
during shipping." (R.T.5391)
Dr. Tonkyn also admitted that, contrary to the information in the
abstract about the Green I aspect of the study, the DOJ had not in fact
conducted nonhuman specimen studies or certain environmental abuse
studies. (R.T. 5341, 5383, 5394) Dr. Tonkyn further testified that when
he gave his oral presentation, he did not discuss with his audience the
DOJ's finding of a number of atypical and unexplainable results, or the
finding of excessive stutter peaks, including one in excess of 20
percent, but he did testify that,
"when I wrote that abstract in August of ‘97, we had been
planning on validating Profiler I. But in between that time, we changed
our minds and decided to validate the Profiler Plus, as that had just
come out. And at the meeting before I started my presentation, I made an
announcement that what was in the proceedings was not exactly correct,
in that we were going to present validation of Profiler Plus, not
Profiler I." (R.T. 5382)
He was adamant that "when I gave the presentation, I prefaced my
prefaced my presentation with the fact that it was not correct; we were
not going to be talking about validation of Profiler I, but Profiler
Plus." (R.T. 5469) He also testified that,
"when I gave the presentation, I stated that the developmental
validation work that we did or experiments that are covered under
developmental validation guidelines of TWGDAM were not complete, and
they weren't the same for both the Green I and Profiler Plus, and I had
listed which ones had been done with which system." (R.T. 5469)
He also testified with respect to stutter that,
"I believe what I stated during my talk was that we were in the
process of collecting stutter data for all of the loci and that we had
not found any stutter peaks that differed greatly from what Perkin Elmer
had listed in their User's Manual, so we might have seen some that were
just slightly by a percentage or two higher than the cut-offs, and that
we were continuing to collect data and we possibly might modify our
stutter cut-offs based on our own findings."
(R.T. 5386)
Unless, therefore, Perkin Elmer employee Holt is an incredibly bad
listener, she could not possibly have heard Tonkyn attest to stutter
guidelines or "experiments to address all the TWGDAM
guidelines," and she certainly could not have heard him talk or
present data on a completed validation study for Profiler I, since no
such study was ever conducted and the audience was so informed. Dr.
Holt's confident testimony to the contrary is compelling evidence of her
bias and her tendency to distort the truth to fit her own (or her
company's) goals.
In the face of this record, the People made the audacious claim in Bokin
that the "widespread use of STRs in forensic casework attests to
the general acceptance of STRs because each laboratory must first,
before commencing casework: (1) generally rely upon the published
Developmental Validation studies, and (2) specifically test the
reliability of the typing kit through Internal Validation studies."
(People's Post Hearing Brief, p. 6) Of course, if the solitary
developmental validation study has not been published or if it is based
on false information, or if the internal validation work relates to
different protocols or is of such a shoddy nature that it cannot
reasonably be relied upon for any purpose, then the logic of the
People's position crumbles.
In the end, the Court in Bokin was inevitably lead by the People's own
witnesses in a giant circle which begins and ends with Perkin Elmer's
so-called validation study and the refusal of the company to disclose
the data underlying it, coupled with an absence of any other relevant
validation studies, and the admission of the People's own witness that
validation specific to the kit must be performed as a prerequisite to
reliable use.
All the defense experts were of the opinion that without the publication
and critical peer review of Perkin Elmer's Green I data and the
appearance of additional extensive validation, there can not be, and has
not been, any general acceptance of the reliability of the Green I kit
in the molecular biology or forensic science communities. The opinions
of these experts is persuasive, particularly in light of a number of
reliability problems which have already surfaced in the brief history of
forensic STR use, including, most recently, the inconsistency between
Perkin Elmer and Promega multiplex STR test kit results discovered by
NIST's Dr. Kline and the New York Medical Examiner in 1997 (Defense
Exhibit G). As Exhibit G indicates, Perkin Elmer has attempted to
downplay this problem, but even the company admits that
"non-concordant results will always be a possibility when the same
DNA sample is amplified using different pairs of primers." (Id.)
But if that is the case, then STR multiplex typing kits have no place in
an adversary system in which each side would be free to choose competing
kits that produce diametrically opposite results and to then make the
claim, based on Dr. Holt's and Perkin Elmer's theory, that both kits are
reliable. As a group of Canadian forensic researchers has observed in
the 1997 Promega Proceeding abstract put into evidence by the People, it
is a fundamental tenet of even the forensic practitioners of DNA typing
that "[c]leary, for the PCR-based STR loci, the type at a given
locus for a given individual should not change as a function of the test
method as was the case in the enzyme dependent RFLP test."
(People's Exhibit 37, p 154; R.T. 180) Even Dr. Holt subscribed to this
obvious principle. (R.T. 181) Yet there is every indication, by Perkin
Elmer's own admission in Exhibit G, that the problem, apparently caused
by a primer binding site mutation, is inevitable and will reoccur in the
future.
There are also a whole range of other reliability problems ably outlined
in Defendant's B70 in Bokin, the NIST survey, Validation Studies
Performed on STR Systems. As there outlined, these problems include:
species specificity problems (the one study conducted to date,
Defendant's Exhibit 80 and People's Tab 20, was conducted on Promega's
silver stain platform), mixture interpretation problems (see, Watts v.
State (1999) __ Miss.__ , 1999 WL 33867 (noting expert testimony
cautioning against extensive reliance upon mixed PCR DNA samples and
stating "[g]iven that evidence, one would have to question the
reliability of any statistical evidence that might be derived therefrom."),
contamination problems stemming from the extreme sensitivity of PCR,
stutter band problems, null alleles, variant alleles, nontemplate
addition of nucleotides, and possible linkage of STR systems to genetic
disease. Ironically, the People included the NIST survey as one of their
prime exhibits (Tab 10) in Bokin, and they cite it for the proposition
that "the articles in peer-reviewed journals unanimously declare
that STRs are forensically reliable." (People's Brief, p. 5) This
is a gross overstatement. The preface to the NIST survey, written in
June 1998, states: "Before a new STR system or STR multiplex may be
routinely employed in human identity testing it should be extensively
validated to insure reliability of results. We have included below many
of the validation studies which have appeared in the literature. We hope
this material will help future researchers design their validation
studies and allow those who use common STR systems to review what
studies have been performed previously." One of the specific
statements in the survey is that "[s]purious PCR products may arise
from non-templated nucleotide addition of adenine or from stutter bands
and thus provide contaminating peaks in an electropherogram."
Clearly, the thrust of the survey is that we are at the beginning of the
validation process for STR systems, not at the end as the People would
have the Court believe. Significantly, not one of the forty six studies
cited in the NIST survey deals specifically with the Green I or Peofiler
Plus kits using capillary electrophoresis and the 310 Genetic Analyzer.
More to the point in Mr. Nawi's case, at least one study conducted bvy
the Department of Defense DNA Registry has concluded that successful
typing of "aged" (8 year old) fingernails is not possible.
See, The Extraction of DNA From Human Nail Material, Proceedings From
the Seventh International Symposium on Human identification (Promega
Corp. 1996) p. 133, attached hereto as Exhibit F.
If Judge Dondero's experience with the disclosure of the data underlying
the DOJ abstract proves anything, it is, in the words of the first NRC
report, that "[i]f a new DNA typing method (or a substantial
variation on an existing one) is to be used in court, publication and
scientific scrutiny are very important." NRC I at 56
"Extensive empirical characterization must be undertaken. Results
must be published in appropriate scientific journals. Publication is the
mechanism that initiates the process of scientific confirmation and
eventual acceptance or rejection of a method." (Id.) ( emphasis
added)
The defense at this stage is not complaining of a failure of discovery
but rather a failure to sustain the burden of proof on the prong one
issue. The court in People v. Axell (1991) 235 Cal. App. 3d 836, 856,
citing State v. Schwartz (Minn. 1989) 447 N.W. 2d 422, declared that
"admissibility of specific test results in a particular case hinges
on the laboratory's compliance with appropriate standards and controls,
and the availability of their testing data and results." The
specific RFLP test results were admissible in Axell, in part because
"Cellmark did comply in this case with defense requests for
discovery and defense experts visited Cellmark's laboratory." Id. p
857 See also, People v. Barney (1992) 8 Cal. App. 4th 798, 813
(contention that general scientific acceptance is precluded by the
failure of the FBI and Cellmark to publish in peer review journals and
otherwise share their data and methodology is foreclosed because
"there have been numerous published articles on DNA analysis as
performed by the FBI, of which we have taken judicial notice.")
By contrast, in State v. Schwartz, supra at 427, the Minnesota Supreme
Court held as follows:
Even if a laboratory has followed reliable procedures to ensure accurate
test results, constitutional concerns may prevent the admissibility of
such evidence. The fair trial and due process rights are implicated when
data relied upon by a laboratory in performing tests are not available
to the opposing party for review and cross examination....
Access to laboratory information generally is significant for another
reason. The validity of testing procedures and principles is assessed in
the scientific community by publishing the data in peer review journals.
The TWGDAM, FBI and CACLD standards stress that publication of a
laboratory's work product and data used in DNA analysis, as well as
independent replication and validation studies, are essential
prerequisites to reliability. Efforts to assess the reliability of the
commercial laboratories' methodology consequently have been hindered
because this information has not yet been made fully available. For
example, Cellmark has not yet published data regarding its methodology
and its probes are only selectively available.
While we agree with the trial court that forensic DNA typing has gained
general acceptance in the scientific community, we hold that
admissibility of specific test results in a particular case hinges on
the laboratory's compliance with appropriate standards and controls, and
the availability of their testing data and results. ... Because the
laboratory in this case did not comport with these guidelines, the test
results lack foundational adequacy and, without more, are thus
inadmissible.
The same result should follow here. The People and Perkin Elmer
cannot in fairness be allowed to have their witnesses cite Perkin
Elmer's study as the sole source of developmental validation and then
turn around and ask the Court to uphold a trade secret privilege as to
the underlying data. It is a matter not only of fairness, but of general
acceptance as well, since all of the defendant's experts, as well as the
NRC reports and the TWGDAM and DAB Guidelines are of one mind in
declaring that particular DNA methods are generally accepted as reliable
within the broad scientific community if, but only if, the methods have
been subjected to extensive validation and peer review. Judge Dondero's
ruling could not be clearer on this point: " the sum of this record
is that the government has not provided the Court with enough to
conclude that STRs identified by using Green One with ABD Genetic
Analyzer 310 satisfies TWGDAM requirements pertaining to developmental
validation. There was no discovery of developmental validation conducted
by Perkin-Elmer. The exhibits referenced by the prosecution in its post
hearing brief deal with peer review of other methods involving STR
identifications." Exhibit A at 9
B. THE DNA EVIDENCE IS INADMISSIBLE BECAUSE THERE ARE NO GENERALLY
ACCEPTED STATISTICAL METHODS THAT ADDRESS BOTH THE PROBABILITY OF A
COINCIDENTAL MATCH BETWEEN TWO PEOPLE WHO SHARE COMMON GENETIC
CHARACTERISTICS AND THE PROBABILITY THAT A MATCH WOULD MISTAKENLY BE
REPORTED DUE TO LABORATORY ERROR.
1. The Failure To Account For Lab Error Rate
Evidence of a DNA "match" between two samples is impossible
to evaluate without reliable information on the likelihood that a match
would be declared if the samples are from different individuals. Most
commentators consider the ability to express this probability to be
crucial to the admissibility of DNA-derived evidence: "without
being informed of such background statistics, the jury is left to its
own speculations." McCormick, Evidence, 655 (Cleary, Ed.).
As indicated at the outset of this opposition, the need for background
statistics to show the meaning of a DNA match is firmly established in
California law. People v. Venegas 18 Cal 4th at 82 ( " A
determination that the DNA profile of an evidentiary sample matches the
profile of a suspect establishes that the two profiles are consistent,
but the determination would be of little significance if the evidentiary
profile also matched that of many or most other human beings."). It
is also established that the statistical calculation phase of DNA
analysis "requires Kelly screening to assure both that the
statistical methodology used is generally accepted in the scientific
community, and that the calculations in the particular case followed
correct scientific procedures." People v. Soto , 21 Cal. 4th at
519.
Since DNA testing methods have been employed in courts, the issue of
statistics, the calculation of genotype frequencies and the proper use
of such statistics in the court room has been an area of contention and
debate. See generally, Symposium: The Evaluation Of Forensic DNA
Evidence, 37 Jurimetrics J. 395 (1997); Devlin, et al., Comments on the
Statistical Aspects of the NRC's Report on DNA Typing, (1994) Journal of
Forensic Sciences, 39(1):28.
Issues surrounding the use of statistical analysis and population
genetics in the forensic arena involve the manner in which a particular
database is constructed and the underlying assumptions in the
statistical analysis. In order for the "product" or
"multiplication" rule, which is used in PCR based testing, to
be valid the each gene that is used must be independent of the other.
This means that there is no actual linkage between two genes that cause
them to be inherited together. One common example of such a linkage is
hair and eye color. If one were to take the frequency of blond hair and
then separately determine the frequency of blue eyes, and multiply them
together as is done with the PM and DQ tests, the resulting frequency of
blue eyed-blonde haired individuals in the population would be much
smaller than the actual occurrence in the population. Blond hair may
occur in 10% (0.10) of the population and blue eyes in 15% (0.15) of the
population. The resulting frequency would be 0.10 x 0.15 = 0.015 or
1.5%. This number however, fails to account for the fact that the most
people with blond hair have blue eyes and therefore, the actual
frequency for blond hair and blue eyed individuals would be much higher
than calculated. This type of linkage or lack of independence is often
called substructure. The issue of whether substructures actually occurs
in human populations is subject to much debate that has yet to be
resolved for PCR based systems.
In its second report, the NRC devoted two chapters to the issue of
statistical analysis of DNA evidence. NRC 2, at 89-166. Soon after the
release of the second NRC report, a symposium was held and papers
published from that symposium on many issues, chief among them the
recommendations of the NRC's second report. See, D.H. Kaye, Dna, Nas,
Nrc, Dab, Rflp, Pcr, And More: an Introduction to The Symposium of The
1996 Nrc Report on Forensic Dna Evidence, Jurimetrics, (Summer 1997)
37:395.
In particular, some commentators recommend using error rates taken from
proficiency tests conducted by laboratories to "correct" the
genotype frequencies calculated in a specific case. Johnathan Koehler,
Why Dna Likelihood Ratios Should Account For Error (Even When a National
Research Council Report Says They Should Not), Jurimetrics, (Summer
1997) 37:425. Others argued that error rates of laboratories be
presented to the jury so that the jury can assess the weight of the
genotype frequencies in light of the relatively high error rates most
laboratories have. William C. Thompson, Accepting Lower Standards: The
Nation Research Council's Second Report on Forensic Evidence,
Jurimetrics, (Summer 1997) 37:405.
Finally, information has recently surfaced that a database used for the
DQ Alpha + polymarker test kit that was published in a peer reviewed
journal and which is used by forensic laboratories to calculate genotype
frequencies was altered to conform the database to the requirements of
Hardy-Weinberg equilibrium. Had the database not been altered, the
database would not have fit within the Hardy-Weinberg assumption which
underlies the multiplication rule. This result calls into question all
of the databases currently in use for forensic testing.
In the midst of these controversies, some consensus has been reached. It
is now broadly recognized that a false "match" between samples
can occur in two ways.
Interpretation of DNA typing results depends not only on population
genetics, but also on laboratory error. Two samples might show the same
DNA pattern for two reasons: two persons have the same genotype at the
loci studied, or the laboratory has made an error in sample handling,
procedure, or interpretation.
NRC 1 at 88.
Thus, to evaluate DNA evidence, the jury needs statistics that address
the probability of both events that could cause a false match. To
provide statistics that reflect the probability of one event that could
cause an innocent person to match, and not the other, would leave the
jury to speculate about the meaning of DNA evidence.
Especially for a technology with high discriminatory power, such as DNA
typing, laboratory error rates must be continually estimated in blind
proficiency testing and must be disclosed to juries. For example,
suppose the chance of a match due to two persons' having the same
pattern were 1 in 1,000,000, but the laboratory had made one error in
500 tests. The jury should be told both results; both facts are relevant
to a jury's determination.
NRC 1 at 89.
The same point was made by Professor Daniel Hartl in an expert's report
filed in U.S. v. Yee. Although his criticism was directed specifically
at the FBI's DNA testing procedures, his point applies to any PCR-based
DNA test:
Once the FBI has declared a match, they treat this declaration as if
there were no operator error and no measurement error. The technology
simply is not up to this standard....In my judgment, experimental error
will turn out to be a very significant term, and perhaps the dominant
term, in any valid estimation of the true probability of a match, and
any statistical calculation that fails to take this into account is
simply meaningless.
Daniel L. Hartl, Expert's Report in the Case of United States v. Yee,
at 4, U.S. v. Bonds, 12 F.3d 540 (6th Cir. 1993), aff'g U.S. v. Yee, 134
F.R.D. 161 (N.D. Ohio 1991) [emphasis added].
Other experts have recently echoed this conclusion:
Statisticians and geneticists involved in the controversy over DNA
testing have understandably been fascinated by and mostly written on
disputes regarding the statistical and genetic issues that DNA
identification raise, but laboratory error places the most serious
limits on the evidentiary import of reported DNA matches. If justice is
the mutual goal of those involved in the debates over DNA
identifications--and I believe it is everyone's concern--the possibility
of error must be honestly faced, and it must be incorporated into
estimates of the incriminatory power of DNA matches.
Richard Lempert, Comment: Theory and Practice in DNA Fingerprinting, 9
Statistical Science 255, 257 (1994).Accord, Laurence D. Mueller, The DNA
Controversy And NRC II, in Statistical Methods in the Health
Sciences:Genetics, ME Halloran & S. Geiser (eds., 1999) pp.17-18;
Jonathan Koehler, Why DNA Likelihood Ratios Should Account For Error
(Even When A National Research Council report Says they Should Not) 37
Jurimetrics 425(1997), R. Lempert, After The DNA Wars: Skirmishing With
NRC II, 37 Jurimetrics 439(1997); William Thompson, Accepting Lower
Standards: The National Research Council's Second Report On Forensic DNA
Evidence, 37 Jurimetrics J. 405-424 (1997); David Balding, Errors and
Misunderstanding In the Second NRC Report, 37 Jurimetrics J. 469 (1997).
In Bokin,the Court has before it the consensus statement of twenty seven
prominent statisticians and population genetists attesting to the fact
that lab error rates must be accounted for in the statistical
computation. ( Bokin Exhibit 105).
Finally, the courts are beginning to heed the well-reasoned statistical
arguments of these experts.With respect to Dr. Koehler, who has written
most prolifically on the issue, it is significant that the Mississipi
Supreme Court very recently declared in Watts v. State (1999) __ Miss.__
, 1999WL 33867 that "as... Dr. Koehler's articles sugges[t], the
introduction of statistical evidence can be meaningless without any
evidence of the testing laboratory's error rate." The Court also
described Dr. Koehler's standing in the scientific community: " Dr.
Koehler, adjunct professor at the University of Texas School of Law and
assistant professor at the University's Buisness, is one of the leading
authorities on the use (and misuse) of statistics in the presentation of
DNA evidence." Id at n. 2
When DNA evidence was first introduced in the courtroom, the primary
concern was the likelihood of a coincidental match. Little attention was
devoted to the likelihood of false positives, perhaps because there was
a widespread misperception that false positives are impossible in RFLP-based
DNA tests. This misperception was generated and sustained by
self-serving claims of DNA test promoters that DNA tests are infallible,
fail-safe, and error-free. Professor Jonathan Koehler has suggested that
DNA test promoters "engaged in a sinister semantic game" in
which they were able to issue misleading denials of the possibility that
a DNA test could make an "error" by excluding consideration of
human error in administering or interpreting the test. Koehler, Error
and Exaggeration, supra, at 24. Needless to say, the effort to
distinguish "human error" from "test error" is
pointless and misleading when humans are necessarily involved in
administration and interpretation of the test, when occasional human
errors are inevitable, and when it is necessary to know the overall rate
of error (from whatever cause) to evaluate the test results. "For
juries it is of little significance what causes an innocent person to
match, what matters is how often such matches might be expected."
Laurence Mueller, The Use of DNA Typing in Forensic Science, 3
Accountability in Research 55, 56 (1993); see also Thompson, Lessons,
supra, at 92.
In any case, the potential for false positives due to laboratory error
in DNA testing is now beyond dispute. "Laboratory errors happen,
even in the best laboratories and even when the analyst is certain that
every precaution against error was taken." NRC 1 at 88-89. See also
Koehler, DNA Matches and Statistics, 76 Judicature 222, 229 ("[B]ased
on the little evidence available to date, a reasonable estimate of the
false positive error rate is 1-4 percent."); Koehler, Error and
Exaggeration, supra, at 26 (proficiency testing shows error rate of
1-4%); Donald Berry, Comment, 9 Stat. Sci. 252, 253 (1994)("Only
the frequency and type of errors are at issue."); R.C. Lewontin,
Comment: The Use of DNA Profiles in Forensic Contexts, 9 Stat. Sci. 259
(1994)(discussing sources of error); William C. Thompson, Comment, 9
Stat. Sci. 263, 265 (1994)(discussing data on laboratory error); Cf. Dan
L. Burk, DNA Identification: Possibilities and Pitfalls Revisited, 31
Jurimetics 53, 80 ("Bald statements or broad hints that DNA testing
is infallible...are not only irresponsible, they border on scientific
fraud").
Indeed, most experts have long believed that having an accurate estimate
of the false positive rate is more important than having an accurate
estimate of the probability of a coincidental match because the rate of
false positives is likely to be much greater than the rate of
coincidental matches, at least for RFLP-based tests. Paul J. Hagerman,
DNA Typing in the Forensic Arena, 47 Am.J.Hum.Genet. 876 (high false
positive rate makes probability of coincidental match irrelevant);
Richard Lempert, Some Caveats Concerning DNA As Criminal Identification
Evidence: With Thanks to the Reverend Bayes, 13 Cardozo L.Rev 303, 325
(the probability of a coincidental match between people who have the
same DNA profile "is usually dwarfed by the probability of a false
positive error"); Mueller, The Use of DNA Typing in Forensic
Science, supra, at 58 (exact probability of a coincidental match
"should hardly matter" to jury given much greater likelihood
of false positive); Richard Ostrowski & Daniel Krane, Unresolved
Issues in Forensic Use of DNA Profiling, 3 Accountability in Research 47
(1993).
A central premise of California appellate cases is that evidence of a
DNA match is inadmissible unless accompanied by statistics that can tell
the trier of fact what the match means. Venegas, Axell, Barney, and
Wallace all recognize this principle. Now that the scientific community
has recognized that error rates must be taken into account in order to
make a meaningful evaluation of DNA evidence, the logic of those cases
requires that juries be given statistics on the probability of
laboratory error; without such statistics, evidence of a DNA match is
inadmissible because it is impossible to evaluate.
Venegas expressly left this issue open (18 Cal. 4th at 95 n. 42), and
Soto does not mention the issue. The 1996 NRC Report recommends without
persuasive analysis that the issue should be simply ignored see, Richard
Lempert, After the DNA Wars: Skirmishing With NRC II, (1997) 37
Jurimetrics J. 439. But it would be absurd for courts to insist on valid
quantitative estimates of the probability of a coincidental match,
without also requiring valid estimates of the rate of false positives
due to laboratory error, when the scientific community has determined
that the latter is more important than the former to a rational
evaluation of DNA evidence. If DNA evidence is "meaningless"
without statistical estimates of the probability of a coincidental
match, it is also "meaningless" without statistical estimates
of the probability of a false positive.
Professor Eric Lander, one of the earliest and most influential
scientific commentators on DNA evidence, and a member of the 1992 NRC
panel, explained the matter succinctly:
it is simply crazy and scientifically unacceptable to agonize over the
exact population frequencies, which might be one in a million, or one in
a hundred thousand, or one in ten thousand for the frequency of a
genotype in a population, and yet not have actual data for the accuracy,
the proficiency of a laboratory's handling of samples...[T]he scientific
acceptability of DNA evidence depends on the proficiency of a laboratory
being tested such that one can know what the error rate is likely to be,
or at least have an upper bound on that error rate.
Prof. Eric Lander, Testimony as a Court's Witness in U.S. v. Porter,
District of Columbia Crim. Docket 3F-6277-89, p. 46, July 28, 1994. In
Soto, the Supreme Court placed great weight on the views of Dr. Lander
in rejecting a challenge to the effect of population substructuring on
RFLP statistical computations. 21 Cal.4th at 538
At this point, there is broad agreement among scientists with the NRC's
position on error rates: DNA evidence cannot be evaluated without
knowing the rate of false positives due to laboratory error, error rates
must therefore be estimated and these estimates must be disclosed to
juries. There is some disagreement, however, about the best way to
present error rate statistics.
The NRC Report recommends that jurors be given two numbers, one
indicating the frequency of matching genotypes and the other indicating
the rate of laboratory error. "The jury should be told both
results; both facts are relevant to a jury's determination." NRC 1
at 89.
Other scientists believe that the error rate and frequency statistics
should be combined into a single number that reflects the overall
likelihood that the laboratory would declare a match between samples
from different people. For example, Professor Hagerman suggested that
"[r]esults of DNA typing should always be reported as the sum of
the laboratory error rate and the estimated frequency of recurrent band
patterns in the relevant population." Hagerman, DNA Typing in the
Forensic Arena, 47 Am. J. Hum. Genet. 876 (1990).
Some experts have gone so far as to suggest that jurors be told only the
false positive rate; they reason that the probability of a false
positive is so much greater than the probability of a coincidental match
(at least for multi-locus RFLP matches) that the latter probability has
little bearing on the value of the evidence. For example, if the
probability of a coincidental match were .000001 (one chance in one
million), and the probability of a false positive were .01 (one chance
in one hundred), then the overall probability of a match between samples
from different people would be approximately .010001, a number that
conveniently rounds off to .01 (one in 100). So why not just tell jurors
the false positive rate and avoid the risk that they will be confused or
unduly swayed by an impressive number (one in one million) that has
little meaning or value relative to the false positive rate?
The rate of false positives defines a practical lower bound on the
probability of a match, and probability estimates based on population
data that are smaller than the false-positive rate should be
disregarded.
R.C. Lewontin & Daniel Hartl, Population Genetics in Forensic DNA
Typing, 254 Science 1745, 1749 (1991).
Professor Richard Lempert specifically cites the danger of confusion and
prejudice as a reason for presenting only the error rate statistic in
cases where the probability of a false positive greatly exceeds the
probability of a coincidental match.
jurors provided with a laboratory's false positive rate and with
information about the likelihood, assuming no testing error, of a match
if the DNA evidence was not the defendant's, are likely to be hopelessly
confused about the weight to accord the testimony because ordinary
people are not very good at working with conditional probabilities.
Thus, jurors ordinarily should receive only the laboratory's false
positive rate as an estimate of the likelihood that the evidence DNA did
not come from the defendant.
Lempert, Caveats, supra, at 325.
Another issue that is being debated is what to tell jurors about the
error rate of a laboratory that has performed relatively few blind
proficiency tests and therefore has limited data on which to base an
error rate estimate. Suppose a laboratory had made no false positive
errors, but had participated in only ten proficiency tests in which a
false positive might have occurred. It would obviously be quite
misleading to tell the jury that the laboratory's error rate is zero. If
the true false positive rate were one in 100, or one in 50, or even one
in 10, there is a very good chance an error would fail to occur in the
first ten trials.
The best solution to this problem is to use a statistical device known
as a confidence interval: the number presented to the jury is not the
actual number of errors, but a number that can be said with a given
degree of confidence (by convention 95% or 99%) to be less than or equal
to the true rate of error. For example, if a laboratory had completed
1000 tests without error, one could say, with 95% confidence, that the
laboratory's true rate of error is less than or equal to .003 (three in
1000), because there is a 95% chance that at least one error would have
been made in 1000 trials if the error rate exceeded .003. Laurence D.
Mueller, The DNA Controversy And NRC II, in Statistical Methods in the
Health Sciences:Genetics, ME Halloran & S. Geiser (eds., 1999)
pp.18(" (L)aboratories should use the upper 95% confidence interval
for their own individual estimates of error when they have a perfect
record, or if the number of tests they have done is very small the
laboratory may use an estimate taken from the industray in
general.");Mueller, The Use of DNA Typing in Forensic Science, 3
Accountability in Research 55 (1993); Saks and Koehler, What DNA
'Fingerprinting' Can Teach the Law About the Rest of Forensic Science,
13 Cardozo L. Rev. 361-372, 369-70; Koehler, DNA Matches and Statistics:
Important Questions, Surprising Answers, 76 Judicature 222, 228 (1993).
The advantage of the confidence interval approach is that it is
conservative. It minimizes the risk that the number reported to the jury
will understate the true error rate and thereby be detrimental to the
interests of the defendant. There is only a five percent chance, for
example, that a number reported with 95% confidence would understate the
true error rate. By contrast, there is a very good chance that the
actual frequency of errors in a limited number of tests will understate
the true error rate, to the detriment of the defendant. Basic principles
of fairness and due process therefore require the use of the confidence
interval approach.
Continuing debate is to be expected over how best to determine the rate
of laboratory error and how best to present error rate data to the jury.
The continuing scientific debate about how to determine and express the
error rate for forensic DNA laboratories reflects a broad consensus that
laboratory error rates are a crucial factor affecting the value of DNA
evidence. The fundamental point recognized by the first NRC Report --
that the error rate must be known in order to evaluate DNA evidence --
is not in dispute. Consequently, unless the People can show that they
have accounted for this issue in some scientifically accepted manner,
this court should hold that all DNA evidence is inadmissible under Kelly
and Venegas.
2. There Is No Generally Accepted Method For Computing The Frequency of
Multi-Locus Profiles.
Based on the report of Allen Keel in this case, it is evident that the
prosecution is seeking to introduce an impressively high random match
statistical probability which combines the six loci in the DQ Alpha test
kit with the one locus in the D1S80 kit and the thirteen STR loci in the
Green I and Profiler-Plus kits. As discussed above, the DQ Alpha +Polymarker
test allows DNA from six different loci (areas of the genome) to be
"amplified" via PCR and then "typed." The D1S80 kit
allows amplification and typing at an additional locus,and the STR loci
adds thirteen more loci to the mix. TheDQ Alpha + polymarker loci have
varying numbers of alleles (types), as does the D1S80 locus and the STR
loci. A "match" on all of the loci in four kits means that two
samples have the same allele (type) at each loci.
Evidence of such a match is meaningless, of course, unless one knows the
probability of finding a match between samples from different people.
This probability depends in part on the likelihood of a false positive
in the tests, which is completely unknown and impossible to estimate in
the absence of rigorous blind proficiency testing (which has not yet
been done). The probability of a match between different people also
depends on the likelihood of a coincidental match, which depends, in
turn, on the rarity of the set of matching alleles. A set of alleles,
from different loci, is called a multi-locus genotype. It is the method
for estimating the frequency of multi-locus genotypes that is discussed
here.
The major problem with the method is that it makes use of the product
rule (multiplication of allele frequencies) without the necessary
demonstration that the alleles are statistically independent. The use of
the product rule assumes that the various alleles are inherited
independently, that there is no reason for persons with a given allele
at one loci to have a preferential probability of having a particular
allele at another loci, and that there is no population substructure.
See NRC 1, Chapter 3. In Soto, the Supreme Court ruled that the debate
regarding the effect of population substructuring on RFLP calculations
had been resolved empirically by "extensive literature in peer
reviewed journals." 21 Cal. 4th at 540 But the debate regarding the
effect of such substructuring on the independence of PCR loci continues
unabated in scientific journals. Indeed, not even the San Francisco
Crime Lab thinks that it can employ the unmodified product rule to PCR
loci. In his report on Mr. Nawi's case, Mr. Keel notes that
" (t)he THO1 and HGBB loci are on chromosome 11 and are
sufficiently close to one another that they are considered to be linked.
However, it is not known whether or not that linkage is sufficiently
close to create associations between these two loci. Similarly, FGA and
GYPA are neighbors on chromosome 4, and D7S8 and D7S820 are neighbors
pon chromosome 7. Until the question of any association between systems
on the same chromosome is addressed in the scientific literature, it is
appropriate to employ frequency information from only one of each of
these neighboring pair of loci.
Report p. 8 n. 8
This reasoning is sound as far as it goes, but it fails to recognize
that all of the loci used in the calculation must be empirically
demonstrated to be be independent. As the NRC I has noted, "[t]he
key question underlying the use of the multiplication rule is whether
actual populations have significant substructure for the loci used for
forensic testing." NRC 1 at 79 [emphasis added]. As defendant's
statistical experts testified in Bokin, without proof of the absence of
substructure for the DQ Alpha +polymarker, D1S80, and particular STR
loci used in this case, the assumption of statistical independence that
underlies the use of the product rule is based entirely on speculation.
In light of this testimony, the two NRC Reports, and the scientific
furor that has arisen over the use of the product rule for RFLP-based
tests, it is clear that the scientific community does not accept its use
for PCR based tests without proof of the independence of the alleles.
Thus, as pointed out above at page 42, nothing in the Soto decision
impacts on the issues in this case. First, it was an RFLP case in which
the Court attached the "greatest significance" to NRC II. 21
Cal. 4th at 539. The NRC II stresses the tentative nature of our current
state of knowledge with respect to PCR-based systems, as well as a
fundamental distinction between RFLP (VNTR) loci and PCR loci.
The lack of information about population substructure also raises severe
doubts about the representativeness of the small convenience samples
that serve as data bases for PCR statistics. There appears to be little
information upon which to judge whether the frequency of particular PCR
alleles in one ethnic group, or one geographic area, is likely to be the
same as in another. Thus, the relevance of small unrepresentative data
bases appears questionable even for estimating the frequency of
individual alleles, let alone estimating the frequency of multi-locus
genotypes.
Assuming it should be applied at all, there can be no doubt that some
modification of the product rule must be made in this case, because the
Supreme Court has already found in People v. Venegas, supra that "
[t]here has been significant scientific controversy over whether
constraints should be imposed on the use of this rule in light of
certain population genetics theories." 18 Cal. 4t at 65 The
existence of this controversy for PCR testing means that there is
presently no general scientific consensus to use the product rule in its
unmodified form for PCR loci. 18 Cal 4th at 84
Moreover, our Supreme Court clearly broke new ground in Venegas, when it
"upheld the trial court's finding of a scientific consensus that
the NRC's ‘modified ceiling' approach or method-used to calculate the
statistical probabilities of a match between the evidentiary samples and
the DNA of an unrelated person chosen at random from the general
population-is forensically reliable, in that from the scientifically
based range of probabilities it selects the figures that most favor the
accused, and therefore cannot furnish a basis on which to invalidate the
admissibility of such evidence on motion of the accused." 18
Cal.4th at 55(emphasis added) Although the term "forensically
reliable" may look and sound like the same statistical artifice
invented by Dr. Bruce Budowle and used by others to downplay the
statistical significance of empirically found departures from Hardy
Weinberg or linkage equilibrium, the concept as used in Venegas is in
fact exactly the opposite of that used by Budowle et al., who employ the
concept to justify numbers that disfavor the accused. As applied to this
case, the principle of forensic reliability as used in Venegas clearly
dictates a modification of the product rule for scientific and due
process reasons.18 Cal. 4th at 84 ( modified ceiling principle must be
adopted because it is "forensically reliable" in that "
it resolves any imprecision in the statistical calculations in a way
that preserves the constitutional presumption of the suspect's
innocence")
The scientific reasons for such a modification were explained in
numerous defense exhibits and in the testimony of defense witnesses Drs.
Laurence Mueller, and Sandy Zabell in Bokin. In general, the exhibits
and the testimony of Drs. Mueller and Zabell established, in the words
of the court in Venegas, that " a figure for the statistical
probabibility of a DNA profile derived from modifying the probabilities
of its individual allele frequencies is accurate only insofar as the
multiplied frequencies are statistically independent from each
other." 18 Cal. 4th at 85 See also, United States v. Shea (D.N.H.
1997) 957 F.Supp. 331, 336, ( "The product rule can be applied
reliably ... only if the estimate of allele frequencies is reasonably
accurate and the conditions in the population approximate what are known
as Hardy-Weinberg equilibrium and linkage equilibrium.")(PCR case)
; United States v. Lowe (D. Mass. 1996) 954 F. Supp. 401,418 ("The
validity of using the product rule in this context depends upon the
independence , or lack of association between the alleles or loci
examined.")(PCR case)
These same witnesses also established, in the words of the court in
State v. Copeland (1996) 130 Wash. Sd 244, 922 P.2d 1304, 1317, that
"potential problems could result from ‘genetic drift' resulting
in small populations having distinct differences, too small a database,
lack of randomness of the samples, and most importantly, lack of a truly
mixed population such that each locus is in Hardy-Weinberg equilibrium
as well as linkage equilibrium." Recent population studies
submitted by the defense, including even studies conducted by
Drs.Budowle and Brenner, two experts heavily relied upon by the People,
have demonstrated a consistent pattern of statistically significant
departures from both HW and linkage equilibium for some of the very
databases used by the San Francisco Police Department.
2. There Is No Generally Accepted Method For Constructing And Using
DNA Databases.
The defense established through the testimony of Mr. Keel in Bokin that
the San Francisco Police Department is currently using an FBI STR
database made up of less than two hundred samples.The People's own
witness, Dr. Holt, testified that an adequate database would require
"hundreds" of database samples. (R.T. 250) See also, People v.
Venegas, 18 Cal. 4th at 64 ("[T]he FBI and Cellmark...and other
forensic laboritories use one or more population databases containing
measurements of the DNA fragments of several hundred persons at each
loci...");United States v. Lowe (D. Mass. 1996) 954 F. Supp.
401,407(quoting NRC II for proposition that an adequate database must be
" at least several hundred persons."); Id. p. 409n. 10( "
the government's experts testified , and Lowe's expert did not dispute,
that a database of several hundred individuals is sufficient.");
Franson v. Micelli (1994) 269 Ill. App. 3d 20, 645 N.E. 2d 404,406
vacated on other grounds, 666 N.W. 2d1188 (Statistical expert
"testified that the ‘Caucasian' data base used by Genetic Design
was too small and that an appropriate random sampling should include at
least 100,000 subjects. He did admit, however, that some statisticians
believe that 5,000 is a valid starting point.") The utilization of
an STR database containing 197 samples undoubtedly introduces
imprecision into the statistical calculation, and defendants experts so
testified.
4. Conclusion
The statistical issues identified by the defense witnesses have lead our
Supreme Court and many other courts to declare the unmodified product
rule inadmissible on a variety of statistical or due process grounds.
See, People v. Venegas, 18 Cal. 4th at 84 ( modified ceiling principle
must be adopted because it is "forensically reliable" in that
" it resolves any imprecision in the statistical calculations in a
way that preserves the constitutional presumption of the suspect's
innocence"), People v. Barney, supra; State v. Carter, 246 Neb.
953, 524 N.W.2d 763, 780-83 (1994) (statistical probability calculation
used in PCR DNA analysis which was based on the Blake database is not
generally accepted within the scientific community) Com. v. Curnin
(1991) 409 Mass. 218, 221-227 , 565 N.E.2d 440, 442-445(there is
currently no general acceptance as to the statistical calculation
process, and thus DNA analysis evidence is inadmissible.); see also
Caldwell v. State (1990) 260 Ga. 278 [393 S.E.2d 436, 444] [laboratory's
statistical evidence inadmissible because state failed to show data
bases were in Hardy- Weinberg equilibrium]; State v. Schwartz (Minn.
1989) 447 N.W.2d 422, 428-429 [exclusion of statistical probability
evidence necessary because of potentially exaggerated impact on jury].)
Under the Kelly standard, the appropriate approach to a modification of
the product rule is to adopt a method of calculation which is both
generally accepted by the scientific community and forensically reliable
within the meaning of Venegas. In this regard, it is significant that in
Venegas, based on the Court's review of both NRC reports, the Court
concluded that " the modified ceiling method ... has been generally
accepted in the relevant scientific community, and that satistical
probability calculations of random matches of DNA profiles computed
under that methodology are admissible under the Kelly test." 18
Cal, 4th at 89
At the very least, therefore, this Court must follow Venegas and rule
that in the absence of the type of statistical showing made in Soto, the
modified ceiling method is one approriate and mandated modification of
the product rule. Further, the Court should decide an issue expressly
left open in Venegas and rule along with the Mississippi Supreme Court
that "as... Dr. Koehler's articles sugges[t], the introduction of
statistical evidence can be meaningless without any evidence of the
testing laboratory's error rate." Watts v. State (1999) __ Miss.__
, 1999WL 33867
The People will no doubt scream that Venegas was an RFLP case. But not
even the critics of Venegas would claim that the unmodified product rule
can be reliably applied in a PCR case. In this regard, it is telling
that the leading proponent of reverse forensic reliability, Dr. Bruce
Budowle, has recently taken the positon that in light of his own
persistent findings of statistically significant departures from HW in
PCR databases, the correct rule is that "[b]y applying the
recommendations of the NRC II Report, these allele frequency data( DQ
A1, PM,D1s80, CSF1PO, TPOX,and THO1) can be used to estimate the rarity
of a multiple locus profile." (Bokin Exhibit 82, People's Tab 26)
See also,United States v. Lowe (D. Mass. 1996) 954 F. Supp. 401,407
("(T)he 10x factor, or confidence interval, is now generally
regarded as a suitable alternative for dealing with the uncertainty that
underlies the statistical approach utilized in DNA analysis.")(PCR
case)
Even under this approach, which is foreclosed by Venegas, the Court
would be obligated to adopt all of the recommendations of NRC II
concerning the product rule, including the use of theta, the 2P rule,
and the 10x factor. Use of all these statistical devices would be
necessary in order to adjust for the several sources of statistical
error outlined in NRC II and by the witnesses in Bokin. Although there
might be some room for debate on whether adjustments like the 2P rule
should be applied to the particular loci at issue here, the defendants
would be required to be given the benefit of the doubt on this issue
pursuant to the principle of forensic reliability established in Venegas.
But again, that same principle mandates use of the modified ceiling
approach in lieu of the recommendations of NRC II.
If any distinction at all is to be made between this cases and Venegas,
it is that in view of the substantial departures from HW and linkage
equilibrium found by Dr. Mueller in PCR databases ( see, Laurence D.
Mueller, The DNA Controversy And NRC II, in Statistical Methods in the
Health Sciences:Genetics, ME Halloran & S. Geiser (eds., 1999);
Mueller, L Commentary on Budowle et. al. validation and population
studies of the loci LDLR, GYPA. HBGG,D7S8, and GC(PM loci), and HLA-DQA
(1998) 43 J Forensic Sci 446-447), and in light of his testimony in
Bokin that departures from linkage equilibrium cannot be corrected by
following the recommendations of the NRC, this case is one in which the
product rule cannot be applied at all because there is presently a lack
of scientific consensus on how to account for such disequilibrium in a
PCR case. This result would be fully justified here because in addition
to the testimony of Drs. Mueller and Zabel, the Court in Bokin had
before it the consensus statement of twenty seven prominent
statisticians and population genetists attesting to the fact that "
many [scientists] oppose the product rule,asserting that there are as
yet insufficient data to determine the effect of population substructure
on probabilty calculations." (Defendants Exhibit 105).Dr Mueller,
it should be noted, is the expert that our Supreme Court relied upon
most heavily in Venegas in finding a prong three statistical violation.
In no event, therefore, is the unmodified product rule justified in this
case.
C. FAILURE OF THE TESTING LABORATORIES TO USE A GENERALLY ACCEPTED
METHOD FOR DETERMINING FALSE POSITIVE ERROR RATES RENDERS THE
PROSECUTION'S DNA EVIDENCE INADMISSIBLE.
1. Essential Elements of an Acceptable Method for Error Rate
Determination.
As noted earlier, the scientific community now recognizes that evidence
of a DNA match cannot meaningfully be evaluated without knowing the rate
of laboratory error. Consequently, under the logic of Venegas, Axell,
Barney and Wallace, evidence of a DNA match cannot be admitted without
statistics on the error rate. To comply with the requirements of Kelly,
the method used to determine the error rate must be generally accepted
as reliable within the relevant scientific community.
Scientific commentary makes it clear that, in order to be accepted as
reliable by the scientific community, the method for determining error
rate must involve externally administered blind proficiency testing on
samples that replicate casework. A blind proficiency test is one in
which the analyst is not aware he or she is being tested. The NRC Report
declares that "...laboratory error rates must be continually
estimated in blind proficiency testing." NRC 1 at 89. The
proficiency tests must be "truly representative of case materials
(with respect to sample quality, accompanying description, etc.)."
Id. The NRC notes that "[t]ests based on pure blood samples would
probably underestimate an error rate." Thus, it appears that such
tests are not accepted as a reliable way to determine the error rate of
a forensic test.
Other commentators agree. For example, Professor Jonathan Koehler
declares that, "[t]he best way to measure the rate of false
positive error associated with a laboratory or an individual technician
is through an ongoing series of blind, external proficiency tests
conducted under realistic conditions." Koehler, Error and
Exaggeration, supra, at 28. Others who endorse this position are R.C.
Lewontin, supra, at 260 ("there must be frequent independent and
unannounced inspections and tests"); Mueller, supra, at 57 (noting
with approval the NRC's call for blind proficiency tests); Hagerman,
supra (noting importance of proficiency testing to error rate
determination).
The NRC Committee emphasized the importance of proficiency testing in a
prefatory statement to its report:
We regard the accreditation and proficiency testing of DNA typing
laboratories as essential to the scientific accuracy, reliability, and
acceptability of DNA typing evidence in the future. Laboratories
involved in forensic DNA typing should move quickly to establish quality
assurance programs. After a sufficient time for implementation of
quality assurance programs has passed, courts should view quality
control as necessary for general acceptance.
NRC 1 at x.
In recent testimony, Professor Eric Lander, a member of the NRC
Committee, explained that the Committee did not wish to call a
moratorium on all forensic DNA testing but wanted to see proficiency
testing programs set up as soon as possible because the Committee
members unanimously believed "that proficiency testing is an
essential component of the scientific reliability and acceptability of
[DNA] evidence," U.S. v. Porter, supra, at 43. In saying that
laboratories should move "quickly" to establish such programs,
the Committee chose the term with care "to be faster than with all
deliberate speed, for example," Id. at 80, and to clearly signal
courts that failure to do proficiency testing would not be acceptable
for long.
Lander also explained the reason the Committee demanded blind
proficiency testing: only blind testing gives an adequate measure of the
likelihood of error.
An adequate [method of proficiency testing] must surely be blind. If you
know that you are working on test samples rather than case samples, you
will, even if you don't intend to, be more careful. Thus, open
proficiency testing where the examiner knows that they are being
examined, does not provide an adequate measure of proficiency. Blind
proficiency testing, where samples are worked in normal case flow,
provides a good measure of that.
Id., at 50.
In light of these very powerful pronouncements demanding realistic blind
proficiency testing, it is clear that there are scientists significant
in number and expertise who would accept no less. It would therefore
appear impossible for any method of error rate estimation to meet the
requirements of Kelly unless it incorporated blind proficiency testing
on samples simulating casework.
2. The SFPD Crime Lab Has Not Employed a Generally Accepted Method to
Estimate Lab Error Rate.
The evidence introduced in Bokin showed that the SFPD Crime Lab has not
completed any external blind proficiency testing on realistic samples to
provide a meaningful estimate of its error rate. Further, the evidence
showed that the open proficiency tests conducted to date have been done
in a haphazard way and in violation of established protocol. The true
error rate of the Lab and its analysts is therefore unknown and
unknowable based on currently available data. Consequently, the value of
the DNA evidence them Lab offers is impossible to evaluate and therefore
inadmissible.
CONCLUSION
For the above-stated reasons, defendant Robert Nawi respectfully
requests that the People's Motion to Admit DNA Test Results be denied in
the absence of a Kelly-Frye hearing at which the People are required to
prove all three prongs of the Kelly- Frye test. Defendant further
requests that the Court take judicial notice of the record and ruling in
People v. Bokin and hold that DNA testing using the Green I kit is
inadmissible as a matter of law.
DATED: December 23,1999
MICHAEL N. BURT
Attorneys for Defendant
Robert Nawi
______________________
By: MICHAEL N. BURT
DEPUTY PUBLIC DEFENDER
|