Jack Kelly writer
 |
| |
 |
| |
 |
| |
 |
| |
 |
DNA TAKES THE STAND
(Invention & Technology, Fall 2006)
In July 1986 residents of the English village of Narborough
learned that someone had raped and murdered a 15-year-
old schoolgirl named Dawn Ashworth. It was a second shock
for the bucolic community, which lies about a hundred miles
north of London. An equally harrowing crime had rattled the
area less than three years earlier. Late in 1983 a passerby
had found the violated body of Lynda Mann, also 15. News of
the latest atrocity frightened and outraged local citizens and
left police desperate.
Authorities focused their investigation on a slow-witted
hospital porter of 17, who, under intense interrogation,
confessed to killing Ashworth. He denied any involvement in
Mann’s killing, but prosecutors were convinced he was lying.
They needed to link him conclusively to both crimes.
Two years earlier, Dr. Alec Jeffreys, a professor of genetics
at the nearby University of Leicester, had developed a way to
identify unique chemical attributes in samples of
deoxyribonucleic acid, the DNA that resides in the nucleus of
every cell. His method had already helped resolve an
immigration case, proving that a young boy from Ghana was
indeed the son of a woman in Britain. Though DNA analysis
had yet to be offered as criminal evidence, police asked
Jeffreys to apply his method to the vexing Narborough case.
They waited for weeks while he performed the delicate steps
of extracting DNA from sperm cells found in the girls’ bodies
and from white blood cells drawn from the suspect. He added
what were known as restriction enzymes, chemicals extracted
from bacteria that precisely cut the strands of DNA at
targeted locations. An electric current dragged the fragments
slowly through a porous gel. The gel acted as a sieve,
separating the short ones, which moved more quickly, from
the longer ones.
Next he transferred the fragments to a membrane in order to
stabilize them. He chemically separated the two strands of
the DNA double helix, using a strong alkali. The key step was
to wash the resulting array with a solution containing a probe,
a section of synthetic DNA formulated to bind with a specific
targeted fragment. Each probe contained a radioactive
isotope. When he laid the membrane on a sheet of X-ray film,
the radiation darkened it in the places where the targeted
fragments were concentrated, creating a pattern of fuzzy
bars. By washing the membrane with one probe after
another, Jeffreys created a series of bars that indicated the
length of different types of fragments.
His method was based on “restriction fragment length
polymorphism,” or RFLP (abbreviations abound in the field).
The term referred to the fact that certain identifiable sections
of DNA vary among individuals. They are shorter in some,
longer in others. His technique was designed to isolate (using
the restriction enzymes), separate (through electrophoresis),
and then visually represent (by means of radioactive
probes). When he analyzed several of the variable sections,
he was able to point with virtual certainty to a single
individual. Yet at the end of his rigorous lab work, he thought
he must have made a mistake.
“My first reaction was ‘Oh my God, there is something wrong
with the technology,’” he said later. His test showed that one
man had indeed raped both girls. But, in spite of his
confession, the youth in custody was not the culprit.
Further testing confirmed the results, and the frustrated
police had to release their suspect, who might well have been
convicted had it not been for the DNA evidence. The police
then took the unprecedented step of requesting blood from
more than 4,000 local men. Samples from those whose blood
type matched that of the evidence—about 10 percent—were
subjected to Jeffreys’s DNA analysis. The laborious
procedure took six months and ended in a cul-de-sac. None
of the samples matched the profile of the murderer.
The break in the case came in August 1987, when a woman
heard a colleague from the bakery where she worked admit
that he had impersonated a co-worker in giving police a
blood sample. When the woman reported the conversation to
the police, they picked up the co-worker, Colin Pitchfork.
Jeffreys’s tests determined that this man’s DNA profile
matched the evidence from both murders. Pitchfork
confessed and, in 1988, was sentenced to life in prison.
Jeffreys labeled his procedure “DNA fingerprinting” (it is now
more commonly called DNA profiling or analysis). This first
case highlighted the potential of the technique. It exonerated
an innocent man and helped police make their case against
a guilty one. The mass screening prefigured the data banks
of DNA samples that have today come to include profiles of
millions of persons. But because of the slow and
cumbersome nature of the process, few imagined that within
a decade DNA analysis would become the most important
forensic tool ever invented. After what he termed an
“entertaining diversion” into forensics, Jeffreys returned to
his basic research, noting that “what comes now is
technological refinement, and that’s not my job.” He was later
knighted for his contribution to science.
Investigators have long searched for methods of identifying
culprits beyond simple (and often unreliable) eyewitness
testimony. Three systems of picking an individual from a
population began to emerge in the late nineteenth century.
In the 1880s a French police official named Alphonse
Bertillon took on the problem of reliably identifying individuals
in order to track criminals and distinguish recidivists from
casual offenders. He devised 11 measurements, including
arm span, cheek width, sitting height, and right-ear length.
Combined, they could identify a person with precision. The
system relied on the fact that while two persons may share
one physical trait, the odds of their having identical values for
a range of measurements was slim. Add enough traits and
the chances of a random match were virtually nonexistent.
The approach was considered scientific because it was both
precise and systematic, giving it an edge over verbal
description or photography.
The weakness of Bertillon’s concept was its application. The
measurements were awkward to take, and small errors
diminished the system’s accuracy. Nevertheless, it was used
by many U.S. police departments into the 1920s. By that time
a simpler, faster, and potentially more useful method had
arrived: the fingerprint.
It had long been known that the tips of the fingers contained
patterns of ridges that varied from one person to another. In
1892 the English scientist Francis Galton determined that
these prints were unique to individuals and permanent
through a person’s life. What was more, secretions left on
items a person touched could contain a record of his
fingerprints and later be used to connect him with a crime
scene. Galton devised a way of classifying the various
patterns. His concept, as revised in 1897, is still used today.
Authorities in the United States first put fingerprint
identification to work in 1910 to help convict a Chicago
murderer. The Federal Bureau of Investigation began
consolidating fingerprint records in 1924 and had
accumulated 100 million examples by the end of World War II.
For decades fingerprints reigned as the principal means of
establishing identity.
The third forensic identifier, blood type, became available in
1901. The Austrian researcher Karl Landsteiner discovered
that the reaction of antigens in incompatible types of blood
formed clumps and prevented transfusion. He labeled the
types A, B, AB, and O. Each was a manifestation of a
particular genetic quirk.
Police used blood-type evidence in criminal investigations to
provide exclusionary evidence. That is, if a criminal left
behind type A blood at the scene, suspects with the other
blood types were ruled out. But simply possessing type A did
not necessarily mean a suspect was guilty, since a large
portion of the population shared that trait.
Over the years, investigators added other blood factors,
along with distinctive serum proteins and enzymes, to aid
forensic identification. Combined, they could increase
discrimination. One person in a hundred might possess a
given blood profile. For determining paternity, such evidence
could be convincing. But in a criminal case, blood-type
evidence could only exclude suspects; it was not specific
enough to erase reasonable doubt.
DNA profiling had connections to all three of these older
means of identification. Like blood typing, it was based on
chemical analysis. Like fingerprinting, it relied on a distinct
but purposeless physical feature that was unique to an
individual. Like Bertillon’s method, it compared a range of
attributes in order to diminish the likelihood of a random
match. But by going to the source of human variability, DNA
profiling held the potential to surpass all earlier methods of
identifying a person or matching a suspect to evidence.
Even as British authorities were struggling to solve the
murder of the two teenagers, detectives in Orlando, Florida,
were on the trail of their own serial rapist. He first struck in
May 1986 and had raped 23 women by the end of the year.
In most cases he broke into a home, showed a knife, covered
the victim’s face to deter identification, and stole some
personal item like a driver’s license.
In February 1987 he raped a young mother while her two
children slept in the next room. This time he left behind two
fingerprints. In March police apprehended a prowler in the
area who matched the prints. He was Tommie Lee Andrews,
a 24-year-old warehouse worker.
The fingerprint evidence was likely to convict Andrews for
that crime, but authorities wanted to connect him to at least
one other rape. It would mean the difference between a few
years in jail and life. However, only one victim could identify
him, and her evidence didn’t guarantee a guilty verdict.
Florida prosecutors turned to DNA profiling, its first use in a U.
S. courtroom. Today DNA analysis is routine, but 20 years
ago the scientists who set about connecting DNA drawn from
Andrews’s blood with that from traces of semen found on the
rape victim faced a daunting challenge in trying to pick out,
from inside two complex molecules, the tiny variations that
were unique to an individual.
DNA, called by one researcher the king of molecules, had
long resisted attempts to understand its nature. As early as
1869 the Swiss physician and chemist Friedrich Miescher
had detected a slightly acidic compound in the nuclei of white
blood cells taken from pus. The molecule’s function wasn’t
pinned down until 1944, when Oswald Avery and his
colleagues at New York’s Rockefeller Institute for Medical
Research determined that DNA was the carrier of genetic
information. But scientists continued to puzzle over how this
unwieldy molecule transmitted all the instructions on which
life depended. The breakthrough came in 1953, when the
American researcher James Watson and his British colleague
Francis Crick determined both the structure of DNA and the
mechanism by which it carried information. An alphabet of
simple chemical substructures, they found, was arranged into
genetic words that regulated activity in the cell.
DNA was made up of a prodigiously long string of atoms. If
the coils upon coils of DNA in a single cell were unwound,
they would stretch six feet. We can picture the molecule as a
twisted ladder, the famous double helix. Human DNA consists
of 46 molecules arranged into 23 pairs of chromosomes (one
of each pair is inherited from each parent). The total of six
billion rungs on these ladders carries the coded information
of life. Each rung consists of one of the four chemical
structures referred to as A, T, G, and C, linked to a
complementary structure on the other strand (A to T, G to C),
forming a “base pair.” Only about 2 percent of the base pairs
in a cell’s DNA—the genes—carry information. The rest of
the molecule has no known function (though some function
may yet be discovered) and is often referred to as noncoding
or “junk” DNA.
In the early 1980s Jeffreys, looking for a way to map the
gene that produced an oxygen-carrying protein, made a
discovery that he admits was “quite accidental.” He began to
examine stutters in the DNA sequence, areas in which a
pattern of base pairs repeated itself, most of them in the
“junk” sections of the molecule. He found that they varied
among individuals in the number of times the sequence was
reiterated. At a particular site, a pattern of 28 base pairs
might be repeated 46 times in one individual, 127 times in
another. Because every site occurs on each of a pair of
chromosomes, a person will usually have two different values
for the repeats. In September of 1984 Jeffreys’s chemical
manipulation succeeded in creating a way to separate
segments of different length and visualize them on a
radiograph.
“It was a ‘Eureka!’ moment,” he has said. Jeffreys had found
a way to identify an aspect of the DNA molecule that was
unique to an individual. “We could immediately see the
potential for forensic investigation.”
It was a major discovery. Although DNA governs all the
physical characteristics that make a person unique, turning it
into a forensic tool was no easy matter because 99.9 percent
of human DNA is the same for everyone. It was Jeffreys’s
discovery of a way to identify and measure what became
known as variable number tandem repeats (VNTR) that led to
further progress.
One quality of DNA that did make it useful for forensic
applications was the fact that the entire code of life was
contained in the nucleus of almost every cell in the body.
That meant that if investigators could compared a stain of
semen with the blood of a suspect, a match would be
possible.
In 1987 technicians at Lifecodes, a laboratory in New York,
were attempting to do just that as they analyzed the DNA
samples in the Tommie Lee Andrews case. They used
chemical detergents and protein digesters to free the DNA
molecules from the cells. After concentrating and purifying
the DNA, they subjected it to the painstaking process of
analysis.
What they were attempting was to cut from the long DNA
molecules specific sections that contained patterns of
tandem repeats. The repeating segment, for example, might
be 20 base pairs long. One person could have 12
repetitions; another, 80. The segments were like boxcars on
a train, and the technicians’ goal was to determine the length
between engine and caboose.
The restriction enzyme sliced out the target segment from the
DNA molecule (it also produced many other fragments that
were irrelevant to the test). The electrophoresis process
used agarose gel (derived from seaweed) as a sieve to make
the long segments move more slowly than the short ones.
Technicians transferred the fragments, now aligned by size,
to a nylon membrane and fixed them in place. A strong alkali
solution broke apart the fragile bonds that held the two
strands together.
The DNA probes they used were synthesized sections of
DNA complementary to the targeted area of repeats. They
relied on the basic principle of DNA that the base A joins only
with T and G only with C (the complement of the sequence
GATC is CTAG). When technicians washed a solution
containing the probes across the nylon membrane, the
probes formed chemical bonds with the targeted sections
that were their complements, but not with miscellaneous
fragments of DNA. By using more than one probe, they were
able to determine the length of several different repeating
sites.
Technicians had “labeled” the probes by including a
radioactive isotope in each. In the areas where they were
concentrated, the probes attached to the target DNA
darkened the X-ray film on which the membrane was laid.
The radiograph thus obtained (it resembled the bar codes on
grocery products) showed the positions of the targeted
fragments, which corresponded with their relative lengths.
In the Tommie Lee Andrews case, the bars created by the
evidence DNA and by that taken from the suspect’s blood
were identical. The jury was charged with weighing a great
deal of new science in reaching its verdict, but in February of
1988 it decided that Andrews was indeed a serial rapist. He
received jail terms totaling 115 years.
In 1989 David Vazquez, who had pleaded guilty to a Virginia
rape and murder in order to avoid a death sentence, became
the first American exonerated on the basis of DNA profiling.
In the meantime the FBI had established its own laboratory
devoted to DNA. The age of forensic DNA, it seemed, was
under way.
Yet the technique was still hampered by a number of
drawbacks. RFLP profiling was hard work: meticulous, time-
consuming, and labor-intensive. The expensive tests took
weeks, sometimes months, to perform. The results required
careful analysis by experts. Moreover, investigators needed
a relatively large sample, a bloodstain the size of a dime, to
extract enough DNA to complete the test.
All that was about to change. A discovery by a California
biochemist was already in the process of revolutionizing
forensic DNA analysis. It would soon shake up the entire field
of molecular biology.
In 1983 Kary Mullis had been working with a team at the
Cetus Corporation, a company exploring the emerging field
of biotechnology, to discover the genetic roots of sickle cell
anemia. One Friday evening, driving north from Berkeley
toward his weekend cottage in Mendocino County, Mullis was
struck by an idea so startling that he pulled his car off the
road and scribbled his idea onto an envelope.
“I thought it was an illusion,” he later said. He understood that
DNA is an awkward molecule to work with, “like an unwound
and tangled audiotape on the floor of the car in the dark.” He
knew that to do anything useful with it, scientists had to focus
on manageable segments. The problem was that any one
short segment of the molecule was so infinitesimally small
that it was valueless. Unless …
He imagined a way to harness the natural ability of DNA to
replicate itself in order to copy a segment of the molecule.
Repeat the process over and over, and the number of copies
would increase geometrically, first 2, then 4, 8, 16, and, after
20 cycles, a million. He had come up with a way to “make as
many copies as I wanted of any DNA sequence I chose.”
He stayed up all night performing calculations. In the coming
months he established the basic elements of what he called
polymerase chain reaction (PCR), though the laboratory
techniques would take years to perfect. He won the Nobel
Prize in 1993 for an invention that was welcomed by
biochemists even as it added to the crime-fighting arsenal of
detectives around the world.
Mullis’s technique had three steps. First he heated a sample
of DNA nearly to the boiling point of water in order to
separate its two strands. Next he cooled the mixture and
relied on “primers,” synthetic DNA fragments, to mark off the
unique section of the chain he wanted to replicate. Like
probes, the primers bonded with their complements along the
length of the sample. They acted like the “find” feature on a
computer, locating a particular sequence of bases.
The critical third step was to warm the solution slightly in
order to encourage the polymerase enzyme to add free-
floating G, A, T, and C base elements to the primers, thereby
spelling out a complementary strand of DNA along the
targeted area. When he heated the mixture again, the new
strand broke off from its template, yielding two segments
ready to be copied again. Repeatedly raising and lowering
the temperature of this mixture—a cycle took about five
minutes—kept the multiplication process going.
When used in forensic analysis, PCR produced a sufficient
amount of DNA after about 25 or 30 cycles. Technicians then
proceeded to separate the fragments by length, using
electrophoresis, as in the RFLP method. Technicians also
began to use fluorescent tags or dyes to visualize the
fragments after electrophoresis, eliminating the slower
radioactive indicators.
PCR was faster, easier, and cheaper than the earlier
method. According to the former Connecticut chief criminalist
Henry Lee, PCR was a technology that “vastly expanded the
ability of forensic scientists to type the DNA of an individual.”
PCR became even more useful in 1991, when Thomas
Caskey, a researcher at the Baylor College of Medicine, in
Houston, suggested the use of DNA segments much shorter
than the variable number tandem repeats Jeffreys used.
These short tandem repeats (STR) consisted of repetitions
of patterns that comprised only three to five base pairs. A
string of repetitions might be 50 base pairs long, while the
sections targeted by Jeffreys were made up of thousands of
base pairs.
As with RFLP, an individual would display two variants
(known as alleles) for any given STR site, one inherited from
each parent. Occasionally the number was the same for
each, but usually it differed. So a PCR test of one site would
create two bands. One might show 5 repeats, the other 12.
This made for an easy comparison with the results of the
same test run on an evidence sample.
Because the fragments were short, they produced a clearer
pattern when separated than did the longer fragments used
in RFLP. Results were unambiguous and easier to interpret.
The process also lent itself to automation, an important factor
when investigators set about creating databases
incorporating thousands of samples.
The most crucial advantage of PCR-based analysis was that
investigators could study much smaller samples than they
could using the RFLP technique. Starting with a minute
quantity, only a few cells, they could “amplify” DNA from the
material, creating as much as they needed. This meant that
tiny blood specks, the saliva on a cigarette butt, or a minute
deposit of skin cells on a ski mask could positively identify a
suspect.
Technicians also found PCR ideal for handling old and
degraded samples. The small DNA segments that constituted
the short tandem repeats were less likely than larger ones to
break apart as the DNA aged. This was important. DNA is a
durable chemical, but heat, ultraviolet light, and other factors
can cause it to disintegrate. Traces of evidence are seldom
pristine, so detectives welcomed a method that could handle
broken DNA. They knew there was no danger of a false
finding. Extreme degradation might prevent a profile from
being obtained, but it never created a different profile.
The advantages of the method meant that by the mid-1990s
PCR had begun to replace the original RFLP technique in
many laboratories.
PCR was just finding its forensic legs in February 1993, when
a bomb exploded in the parking garage of the World Trade
Center. The blast killed 6 people and injured more than
1,000. Four days later editors at The New York Times
received a letter claiming responsibility for the attack.
One of the bombers, Mohammed Salameh, was arrested
soon afterward when he went to collect a deposit for the
rental truck that had carried the explosive. Authorities quickly
focused on Salameh’s associate Nidal Ayyad. Included in the
evidence that linked him to the bombing was a trace of saliva
on the envelope flap of the letter received by the Times. The
PCR technique used then was not as discriminating as it
would become later, but experts estimated that only one of
50 people would match the profile found on the envelope,
and Ayyad was one of them. The evidence helped the jury
convict him.
Because the new form of DNA analysis relied on much
shorter repeating segments and therefore lacked the
discriminating power of Jeffreys’s original method,
investigators began to target more STR sites along the
molecule for amplification (thousands are known to exist). To
do so, they simply added primers formulated to attach to
different DNA regions. Two suspects might show the same
pattern for one site, but the chances of a random match
diminished exponentially with each site added. The FBI
eventually settled on the 13 sites that now make up a
standard DNA profile.
The main drawback of PCR was its sensitivity to
contamination. If a few skin cells from a lab worker fell into a
drop of blood that was to be analyzed by the RFLP method,
their DNA would be lost in the much larger quantity of DNA
from the sample itself. But if a contaminant was added to a
tiny trace of evidence intended for PCR analysis, the process
would amplify DNA from the contaminant as well as the
evidence, resulting in a confusing or misleading outcome.
PCR demanded scrupulous evidence handling and
superclean laboratory techniques.
Another difficulty was pinning down exactly how common a
particular pattern might be. DNA analysis never deals in
certainties, only in probabilities. The results indicate that a
particular person’s DNA pattern matches the pattern of an
evidence sample. Statisticians then have to calculate the
chances of a person chosen at random having a similar
match. Is it one in a thousand, one in a million, or one in a
billion?
These numbers are obtained by the product rule. Let’s say
for simplicity that there are 10 possible variants in the
number of repeats at each STR site (the actual number is
usually higher). If the variants are randomly distributed, the
chance of a match at one site is 1 in 10. The chance of
matches at two sites is 1 in 100; at three, 1 in 1,000; at four,
1 in 10,000. Each additional site decreases the likelihood of
a random match by a power of 10. Factoring all 13 STR sites
into the equation, researchers now set the chance of a
random match between two DNA samples at a staggering
one in 575 trillion. When prosecutors explain such numbers
to a jury, DNA evidence becomes very powerful.
In 1994 a case that would make forensic DNA profiling
familiar to millions burst into the headlines. California
prosecutors accused the retired football player Orenthal
James Simpson of stabbing to death his former wife, Nicole
Brown, and her friend Ronald Goldman. DNA became a
central issue in the ensuing trial. The jury sat through what
amounted to a crash course in DNA analysis, and the
technique was mentioned more than 10,000 times in the
transcript.
The analysis produced damning evidence against the
accused. The blood of both victims stained a glove found
outside Simpson’s home. Tests on a stain from Simpson’s
Ford Bronco found that it contained Goldman’s blood. Blood
on a sock found on the floor of Simpson’s bedroom was
identified as Nicole Brown’s. It seemed that the new science
had done an admirable job in unraveling the mystery.
But while the defense did not contest the admissibility of DNA
evidence, Simpson’s “dream team” of lawyers was able to
raise troubling questions about the techniques used to
gather and process the evidence. Much of the testing relied
on the original RFLP technique, which was subject to
laboratory error and interpretation. The newer PCR method,
on the other hand, was acutely susceptible to contamination.
Technicians were notably careless in collecting and handling
samples. One lab worker tested evidence soon after spilling
reference blood drawn from Simpson. These procedural
deficiencies, along with suggestions that Simpson had been
framed by investigators, contributed to the jury’s verdict of
not guilty.
Prosecutors and police realized that scrupulous evidence
handling and rigorous laboratory methods were needed if
DNA evidence was to survive scrutiny. “Based on the O. J.
Simpson case,” the New York City police commissioner
Howard Safir said in 1998, “we had to design a lab and
procedures that are much less subject to challenge.”
In 1996 the federal government began to award grants to
states through its DNA Laboratory Improvement Program.
The funds went for upgrading and standardizing procedures,
and they helped labs switch from RFLP to PCR technology.
That same year the National Research Council gave its
stamp of approval to DNA evidence, declaring that testing
techniques and statistical data had “progressed to the point
where the admissibility of properly collected and analyzed
data should not be in doubt.” Juries were soon routinely
accepting as decisive evidence based on DNA analysis.
In the late 1990s DNA profiling hit the news again, providing
evidence in scandals involving two American Presidents. The
first was almost 200 years old. As early as 1802 Thomas
Jefferson’s enemies had accused him of fathering a child with
his slave Sally Hemings. Proof one way or the other seemed
impossible until 1998. An examination of the DNA of
descendants of Thomas Jefferson’s uncle (Jefferson himself
had no known male heirs) and of that of Hemings’s progeny
suggested that Jefferson had sired at least one of her
children.
Even as it was pointing the finger back into history, DNA
evidence caught in a lie the then current occupant of the
White House, Bill Clinton. Clinton had assured the nation that
he had carried on no sexual relationship with “that woman,”
the White House intern Monica Lewinsky. A DNA profile
ordered by the special prosecutor, Kenneth Starr, proved
otherwise, linking semen stains on Lewinsky’s blue dress to
blood drawn from the President. The irrefutable evidence led
to a shamefaced confession and laid the groundwork for the
first impeachment of an American President in 130 years.
Because of its decisive ability to exclude a suspect, DNA has
always been a powerful tool of exoneration. In 1981 Robert
Clark, 20, was accused of rape, kidnapping, and armed
robbery for a crime that took place in East Atlanta, Georgia.
He proclaimed his innocence, even naming the man from
whom he had bought the victim’s car, which Clark was
assumed to have stolen. But the victim picked him out of a
lineup. He was convicted and sentenced to two life terms plus
20 years.
In 2003 Clark contacted the Innocence Project, an
organization that the lawyers Barry Scheck and Peter Neufeld
had founded in 1992 at New York University’s Benjamin N.
Cardozo School of Law. The nonprofit group provides post-
conviction DNA testing in cases where there is a question of
guilt.
Over the objection of prosecutors, Project lawyers arranged
for DNA analysis in the Clark case. The test proved that
Robert Clark was not the perpetrator of the crime for which
he had been convicted. Instead, it implicated the very man
Clark had named two decades earlier. In December 2005,
after almost 25 years in prison, Clark walked free.
The role of DNA testing in exonerating the wrongfully
convicted has been its most important contribution to justice.
Every mistaken conviction perpetrates a double injustice: an
innocent person locked up, a criminal unpunished. Robert
Clark joined the 175 other defendants who have so far been
cleared by the Innocence Project. This success has spawned
many similar efforts. It has raised questions about the
reliability of other forensic evidence, including confessions
and eyewitness identifications. In 2003 the Illinois governor,
George Ryan, commuted the sentences of all his state’s
death-row inmates after 13 of them were found to be
innocent by DNA testing and other means. Nationwide, DNA
testing alone has cleared more than a dozen persons
wrongfully condemned to death.
Back in 1977, years before DNA profiling was dreamed of, a
robber held up a Richmond, Virginia, business called Shakey’
s Pizza Parlor. The incident turned violent, and the thief
murdered the shop’s owner. In the struggle the assailant left
behind drops of his own blood. With little additional evidence
and no suspects, police placed the case on the list of those
they would probably never solve.
DNA profiling can link a suspect to a crime or exclude him.
But what about those cases where there is no suspect? State
and federal authorities have long kept fingerprint records of
criminals and others so that if a print from an unknown
person is found at a crime scene, they can try to find a
match. They realized that a similar database containing DNA
profiles would allow them to attempt to link evidence from a
crime scene to the profiles of potential suspects.
In 1998 the FBI began a program known as the Combined
DNA Index System, or CODIS, which consolidates DNA
profiles from the states into a national registry. The criteria
for collecting the samples vary by state; many require all
convicted felons and prison inmates to submit a sample. The
PCR technique assigns two figures, indicating the number of
repeats, to each of the 13 DNA sites analyzed. One figure is
a measure of the version of the site inherited from the
mother, the other of that from the father. These 26 numbers
represent a complete, compact profile of a person’s DNA,
allowing for an easily searchable database.
The system has enabled investigators to match crime scene
evidence to the DNA of a potential suspect in more than
30,000 open cases. It has helped solve crimes that are
decades old. By 2006 the FBI had collected 3.2 million
profiles, along with 136,000 evidence samples from unsolved
crimes.
Blood evidence from the Virginia pizza parlor murder was
included in that state’s DNA database in 1996. Then, in
2004, police required a man arrested on a felony gun charge
to provide a blood sample for DNA typing. The two samples
matched. In February 2006 Benjamin Richard Johnson, now
61, was charged with the 29-year-old murder.
A DNA database raises questions of privacy and fairness.
Whose DNA should be included? Only convicted criminals’?
Everyone’s? What should happen to the DNA sample after it
has been analyzed? If a person in a database is a near-
perfect match, should his or her relatives, who might share a
similar profile, become suspects?
Today DNA profiling does not provide any genetic
information about a suspect. Like fingerprinting, it focuses on
a unique but meaningless feature. But the rapid evolution of
genetic science raises the possibility that analysts will soon
be able to infer from DNA physical characteristics like eye
and skin color, approximate height, propensity to illnesses,
and ethnic background. Even a psychological profile is not
out of the question. Is the person prone to anger or violent
action? These possibilities will require society to grapple with
novel ethical questions, but they also hold promise for
effectively using science to bring wrongdoers to justice.
Riding the wave of discoveries in molecular biology and
borrowing the tools developed to map the human genome,
DNA profiling has progressed rapidly over the past decade.
Mechanization has made all aspects of DNA analysis
increasingly automatic. What originally took hours or days
can now be done in minutes. Instead of amplifying and
analyzing a single target section at a time, investigators can
look at all 13 standard sites simultaneously. Instead of
creating bar graphs, they allow a laser detector to profile
segments as the DNA fragments pass down a narrow tube
and let computers analyze the information. Within a few
years, police will be able to analyze evidence and produce a
profile in minutes using hand-held devices right at the scene
of a crime. By tapping into a computerized database, they
may be able to implicate a specific individual almost
immediately.
So powerful is DNA analysis that today a fingerprint may not
only reveal a clue to a perpetrator’s identity in itself but also
yield to investigators a few sloughed skin cells from which
they can extract DNA, develop a profile, and use it to
conclusively nail the culprit.
# # #
(Invention & Technology, Fall 2006)
In July 1986 residents of the English village of Narborough
learned that someone had raped and murdered a 15-year-
old schoolgirl named Dawn Ashworth. It was a second shock
for the bucolic community, which lies about a hundred miles
north of London. An equally harrowing crime had rattled the
area less than three years earlier. Late in 1983 a passerby
had found the violated body of Lynda Mann, also 15. News of
the latest atrocity frightened and outraged local citizens and
left police desperate.
Authorities focused their investigation on a slow-witted
hospital porter of 17, who, under intense interrogation,
confessed to killing Ashworth. He denied any involvement in
Mann’s killing, but prosecutors were convinced he was lying.
They needed to link him conclusively to both crimes.
Two years earlier, Dr. Alec Jeffreys, a professor of genetics
at the nearby University of Leicester, had developed a way to
identify unique chemical attributes in samples of
deoxyribonucleic acid, the DNA that resides in the nucleus of
every cell. His method had already helped resolve an
immigration case, proving that a young boy from Ghana was
indeed the son of a woman in Britain. Though DNA analysis
had yet to be offered as criminal evidence, police asked
Jeffreys to apply his method to the vexing Narborough case.
They waited for weeks while he performed the delicate steps
of extracting DNA from sperm cells found in the girls’ bodies
and from white blood cells drawn from the suspect. He added
what were known as restriction enzymes, chemicals extracted
from bacteria that precisely cut the strands of DNA at
targeted locations. An electric current dragged the fragments
slowly through a porous gel. The gel acted as a sieve,
separating the short ones, which moved more quickly, from
the longer ones.
Next he transferred the fragments to a membrane in order to
stabilize them. He chemically separated the two strands of
the DNA double helix, using a strong alkali. The key step was
to wash the resulting array with a solution containing a probe,
a section of synthetic DNA formulated to bind with a specific
targeted fragment. Each probe contained a radioactive
isotope. When he laid the membrane on a sheet of X-ray film,
the radiation darkened it in the places where the targeted
fragments were concentrated, creating a pattern of fuzzy
bars. By washing the membrane with one probe after
another, Jeffreys created a series of bars that indicated the
length of different types of fragments.
His method was based on “restriction fragment length
polymorphism,” or RFLP (abbreviations abound in the field).
The term referred to the fact that certain identifiable sections
of DNA vary among individuals. They are shorter in some,
longer in others. His technique was designed to isolate (using
the restriction enzymes), separate (through electrophoresis),
and then visually represent (by means of radioactive
probes). When he analyzed several of the variable sections,
he was able to point with virtual certainty to a single
individual. Yet at the end of his rigorous lab work, he thought
he must have made a mistake.
“My first reaction was ‘Oh my God, there is something wrong
with the technology,’” he said later. His test showed that one
man had indeed raped both girls. But, in spite of his
confession, the youth in custody was not the culprit.
Further testing confirmed the results, and the frustrated
police had to release their suspect, who might well have been
convicted had it not been for the DNA evidence. The police
then took the unprecedented step of requesting blood from
more than 4,000 local men. Samples from those whose blood
type matched that of the evidence—about 10 percent—were
subjected to Jeffreys’s DNA analysis. The laborious
procedure took six months and ended in a cul-de-sac. None
of the samples matched the profile of the murderer.
The break in the case came in August 1987, when a woman
heard a colleague from the bakery where she worked admit
that he had impersonated a co-worker in giving police a
blood sample. When the woman reported the conversation to
the police, they picked up the co-worker, Colin Pitchfork.
Jeffreys’s tests determined that this man’s DNA profile
matched the evidence from both murders. Pitchfork
confessed and, in 1988, was sentenced to life in prison.
Jeffreys labeled his procedure “DNA fingerprinting” (it is now
more commonly called DNA profiling or analysis). This first
case highlighted the potential of the technique. It exonerated
an innocent man and helped police make their case against
a guilty one. The mass screening prefigured the data banks
of DNA samples that have today come to include profiles of
millions of persons. But because of the slow and
cumbersome nature of the process, few imagined that within
a decade DNA analysis would become the most important
forensic tool ever invented. After what he termed an
“entertaining diversion” into forensics, Jeffreys returned to
his basic research, noting that “what comes now is
technological refinement, and that’s not my job.” He was later
knighted for his contribution to science.
Investigators have long searched for methods of identifying
culprits beyond simple (and often unreliable) eyewitness
testimony. Three systems of picking an individual from a
population began to emerge in the late nineteenth century.
In the 1880s a French police official named Alphonse
Bertillon took on the problem of reliably identifying individuals
in order to track criminals and distinguish recidivists from
casual offenders. He devised 11 measurements, including
arm span, cheek width, sitting height, and right-ear length.
Combined, they could identify a person with precision. The
system relied on the fact that while two persons may share
one physical trait, the odds of their having identical values for
a range of measurements was slim. Add enough traits and
the chances of a random match were virtually nonexistent.
The approach was considered scientific because it was both
precise and systematic, giving it an edge over verbal
description or photography.
The weakness of Bertillon’s concept was its application. The
measurements were awkward to take, and small errors
diminished the system’s accuracy. Nevertheless, it was used
by many U.S. police departments into the 1920s. By that time
a simpler, faster, and potentially more useful method had
arrived: the fingerprint.
It had long been known that the tips of the fingers contained
patterns of ridges that varied from one person to another. In
1892 the English scientist Francis Galton determined that
these prints were unique to individuals and permanent
through a person’s life. What was more, secretions left on
items a person touched could contain a record of his
fingerprints and later be used to connect him with a crime
scene. Galton devised a way of classifying the various
patterns. His concept, as revised in 1897, is still used today.
Authorities in the United States first put fingerprint
identification to work in 1910 to help convict a Chicago
murderer. The Federal Bureau of Investigation began
consolidating fingerprint records in 1924 and had
accumulated 100 million examples by the end of World War II.
For decades fingerprints reigned as the principal means of
establishing identity.
The third forensic identifier, blood type, became available in
1901. The Austrian researcher Karl Landsteiner discovered
that the reaction of antigens in incompatible types of blood
formed clumps and prevented transfusion. He labeled the
types A, B, AB, and O. Each was a manifestation of a
particular genetic quirk.
Police used blood-type evidence in criminal investigations to
provide exclusionary evidence. That is, if a criminal left
behind type A blood at the scene, suspects with the other
blood types were ruled out. But simply possessing type A did
not necessarily mean a suspect was guilty, since a large
portion of the population shared that trait.
Over the years, investigators added other blood factors,
along with distinctive serum proteins and enzymes, to aid
forensic identification. Combined, they could increase
discrimination. One person in a hundred might possess a
given blood profile. For determining paternity, such evidence
could be convincing. But in a criminal case, blood-type
evidence could only exclude suspects; it was not specific
enough to erase reasonable doubt.
DNA profiling had connections to all three of these older
means of identification. Like blood typing, it was based on
chemical analysis. Like fingerprinting, it relied on a distinct
but purposeless physical feature that was unique to an
individual. Like Bertillon’s method, it compared a range of
attributes in order to diminish the likelihood of a random
match. But by going to the source of human variability, DNA
profiling held the potential to surpass all earlier methods of
identifying a person or matching a suspect to evidence.
Even as British authorities were struggling to solve the
murder of the two teenagers, detectives in Orlando, Florida,
were on the trail of their own serial rapist. He first struck in
May 1986 and had raped 23 women by the end of the year.
In most cases he broke into a home, showed a knife, covered
the victim’s face to deter identification, and stole some
personal item like a driver’s license.
In February 1987 he raped a young mother while her two
children slept in the next room. This time he left behind two
fingerprints. In March police apprehended a prowler in the
area who matched the prints. He was Tommie Lee Andrews,
a 24-year-old warehouse worker.
The fingerprint evidence was likely to convict Andrews for
that crime, but authorities wanted to connect him to at least
one other rape. It would mean the difference between a few
years in jail and life. However, only one victim could identify
him, and her evidence didn’t guarantee a guilty verdict.
Florida prosecutors turned to DNA profiling, its first use in a U.
S. courtroom. Today DNA analysis is routine, but 20 years
ago the scientists who set about connecting DNA drawn from
Andrews’s blood with that from traces of semen found on the
rape victim faced a daunting challenge in trying to pick out,
from inside two complex molecules, the tiny variations that
were unique to an individual.
DNA, called by one researcher the king of molecules, had
long resisted attempts to understand its nature. As early as
1869 the Swiss physician and chemist Friedrich Miescher
had detected a slightly acidic compound in the nuclei of white
blood cells taken from pus. The molecule’s function wasn’t
pinned down until 1944, when Oswald Avery and his
colleagues at New York’s Rockefeller Institute for Medical
Research determined that DNA was the carrier of genetic
information. But scientists continued to puzzle over how this
unwieldy molecule transmitted all the instructions on which
life depended. The breakthrough came in 1953, when the
American researcher James Watson and his British colleague
Francis Crick determined both the structure of DNA and the
mechanism by which it carried information. An alphabet of
simple chemical substructures, they found, was arranged into
genetic words that regulated activity in the cell.
DNA was made up of a prodigiously long string of atoms. If
the coils upon coils of DNA in a single cell were unwound,
they would stretch six feet. We can picture the molecule as a
twisted ladder, the famous double helix. Human DNA consists
of 46 molecules arranged into 23 pairs of chromosomes (one
of each pair is inherited from each parent). The total of six
billion rungs on these ladders carries the coded information
of life. Each rung consists of one of the four chemical
structures referred to as A, T, G, and C, linked to a
complementary structure on the other strand (A to T, G to C),
forming a “base pair.” Only about 2 percent of the base pairs
in a cell’s DNA—the genes—carry information. The rest of
the molecule has no known function (though some function
may yet be discovered) and is often referred to as noncoding
or “junk” DNA.
In the early 1980s Jeffreys, looking for a way to map the
gene that produced an oxygen-carrying protein, made a
discovery that he admits was “quite accidental.” He began to
examine stutters in the DNA sequence, areas in which a
pattern of base pairs repeated itself, most of them in the
“junk” sections of the molecule. He found that they varied
among individuals in the number of times the sequence was
reiterated. At a particular site, a pattern of 28 base pairs
might be repeated 46 times in one individual, 127 times in
another. Because every site occurs on each of a pair of
chromosomes, a person will usually have two different values
for the repeats. In September of 1984 Jeffreys’s chemical
manipulation succeeded in creating a way to separate
segments of different length and visualize them on a
radiograph.
“It was a ‘Eureka!’ moment,” he has said. Jeffreys had found
a way to identify an aspect of the DNA molecule that was
unique to an individual. “We could immediately see the
potential for forensic investigation.”
It was a major discovery. Although DNA governs all the
physical characteristics that make a person unique, turning it
into a forensic tool was no easy matter because 99.9 percent
of human DNA is the same for everyone. It was Jeffreys’s
discovery of a way to identify and measure what became
known as variable number tandem repeats (VNTR) that led to
further progress.
One quality of DNA that did make it useful for forensic
applications was the fact that the entire code of life was
contained in the nucleus of almost every cell in the body.
That meant that if investigators could compared a stain of
semen with the blood of a suspect, a match would be
possible.
In 1987 technicians at Lifecodes, a laboratory in New York,
were attempting to do just that as they analyzed the DNA
samples in the Tommie Lee Andrews case. They used
chemical detergents and protein digesters to free the DNA
molecules from the cells. After concentrating and purifying
the DNA, they subjected it to the painstaking process of
analysis.
What they were attempting was to cut from the long DNA
molecules specific sections that contained patterns of
tandem repeats. The repeating segment, for example, might
be 20 base pairs long. One person could have 12
repetitions; another, 80. The segments were like boxcars on
a train, and the technicians’ goal was to determine the length
between engine and caboose.
The restriction enzyme sliced out the target segment from the
DNA molecule (it also produced many other fragments that
were irrelevant to the test). The electrophoresis process
used agarose gel (derived from seaweed) as a sieve to make
the long segments move more slowly than the short ones.
Technicians transferred the fragments, now aligned by size,
to a nylon membrane and fixed them in place. A strong alkali
solution broke apart the fragile bonds that held the two
strands together.
The DNA probes they used were synthesized sections of
DNA complementary to the targeted area of repeats. They
relied on the basic principle of DNA that the base A joins only
with T and G only with C (the complement of the sequence
GATC is CTAG). When technicians washed a solution
containing the probes across the nylon membrane, the
probes formed chemical bonds with the targeted sections
that were their complements, but not with miscellaneous
fragments of DNA. By using more than one probe, they were
able to determine the length of several different repeating
sites.
Technicians had “labeled” the probes by including a
radioactive isotope in each. In the areas where they were
concentrated, the probes attached to the target DNA
darkened the X-ray film on which the membrane was laid.
The radiograph thus obtained (it resembled the bar codes on
grocery products) showed the positions of the targeted
fragments, which corresponded with their relative lengths.
In the Tommie Lee Andrews case, the bars created by the
evidence DNA and by that taken from the suspect’s blood
were identical. The jury was charged with weighing a great
deal of new science in reaching its verdict, but in February of
1988 it decided that Andrews was indeed a serial rapist. He
received jail terms totaling 115 years.
In 1989 David Vazquez, who had pleaded guilty to a Virginia
rape and murder in order to avoid a death sentence, became
the first American exonerated on the basis of DNA profiling.
In the meantime the FBI had established its own laboratory
devoted to DNA. The age of forensic DNA, it seemed, was
under way.
Yet the technique was still hampered by a number of
drawbacks. RFLP profiling was hard work: meticulous, time-
consuming, and labor-intensive. The expensive tests took
weeks, sometimes months, to perform. The results required
careful analysis by experts. Moreover, investigators needed
a relatively large sample, a bloodstain the size of a dime, to
extract enough DNA to complete the test.
All that was about to change. A discovery by a California
biochemist was already in the process of revolutionizing
forensic DNA analysis. It would soon shake up the entire field
of molecular biology.
In 1983 Kary Mullis had been working with a team at the
Cetus Corporation, a company exploring the emerging field
of biotechnology, to discover the genetic roots of sickle cell
anemia. One Friday evening, driving north from Berkeley
toward his weekend cottage in Mendocino County, Mullis was
struck by an idea so startling that he pulled his car off the
road and scribbled his idea onto an envelope.
“I thought it was an illusion,” he later said. He understood that
DNA is an awkward molecule to work with, “like an unwound
and tangled audiotape on the floor of the car in the dark.” He
knew that to do anything useful with it, scientists had to focus
on manageable segments. The problem was that any one
short segment of the molecule was so infinitesimally small
that it was valueless. Unless …
He imagined a way to harness the natural ability of DNA to
replicate itself in order to copy a segment of the molecule.
Repeat the process over and over, and the number of copies
would increase geometrically, first 2, then 4, 8, 16, and, after
20 cycles, a million. He had come up with a way to “make as
many copies as I wanted of any DNA sequence I chose.”
He stayed up all night performing calculations. In the coming
months he established the basic elements of what he called
polymerase chain reaction (PCR), though the laboratory
techniques would take years to perfect. He won the Nobel
Prize in 1993 for an invention that was welcomed by
biochemists even as it added to the crime-fighting arsenal of
detectives around the world.
Mullis’s technique had three steps. First he heated a sample
of DNA nearly to the boiling point of water in order to
separate its two strands. Next he cooled the mixture and
relied on “primers,” synthetic DNA fragments, to mark off the
unique section of the chain he wanted to replicate. Like
probes, the primers bonded with their complements along the
length of the sample. They acted like the “find” feature on a
computer, locating a particular sequence of bases.
The critical third step was to warm the solution slightly in
order to encourage the polymerase enzyme to add free-
floating G, A, T, and C base elements to the primers, thereby
spelling out a complementary strand of DNA along the
targeted area. When he heated the mixture again, the new
strand broke off from its template, yielding two segments
ready to be copied again. Repeatedly raising and lowering
the temperature of this mixture—a cycle took about five
minutes—kept the multiplication process going.
When used in forensic analysis, PCR produced a sufficient
amount of DNA after about 25 or 30 cycles. Technicians then
proceeded to separate the fragments by length, using
electrophoresis, as in the RFLP method. Technicians also
began to use fluorescent tags or dyes to visualize the
fragments after electrophoresis, eliminating the slower
radioactive indicators.
PCR was faster, easier, and cheaper than the earlier
method. According to the former Connecticut chief criminalist
Henry Lee, PCR was a technology that “vastly expanded the
ability of forensic scientists to type the DNA of an individual.”
PCR became even more useful in 1991, when Thomas
Caskey, a researcher at the Baylor College of Medicine, in
Houston, suggested the use of DNA segments much shorter
than the variable number tandem repeats Jeffreys used.
These short tandem repeats (STR) consisted of repetitions
of patterns that comprised only three to five base pairs. A
string of repetitions might be 50 base pairs long, while the
sections targeted by Jeffreys were made up of thousands of
base pairs.
As with RFLP, an individual would display two variants
(known as alleles) for any given STR site, one inherited from
each parent. Occasionally the number was the same for
each, but usually it differed. So a PCR test of one site would
create two bands. One might show 5 repeats, the other 12.
This made for an easy comparison with the results of the
same test run on an evidence sample.
Because the fragments were short, they produced a clearer
pattern when separated than did the longer fragments used
in RFLP. Results were unambiguous and easier to interpret.
The process also lent itself to automation, an important factor
when investigators set about creating databases
incorporating thousands of samples.
The most crucial advantage of PCR-based analysis was that
investigators could study much smaller samples than they
could using the RFLP technique. Starting with a minute
quantity, only a few cells, they could “amplify” DNA from the
material, creating as much as they needed. This meant that
tiny blood specks, the saliva on a cigarette butt, or a minute
deposit of skin cells on a ski mask could positively identify a
suspect.
Technicians also found PCR ideal for handling old and
degraded samples. The small DNA segments that constituted
the short tandem repeats were less likely than larger ones to
break apart as the DNA aged. This was important. DNA is a
durable chemical, but heat, ultraviolet light, and other factors
can cause it to disintegrate. Traces of evidence are seldom
pristine, so detectives welcomed a method that could handle
broken DNA. They knew there was no danger of a false
finding. Extreme degradation might prevent a profile from
being obtained, but it never created a different profile.
The advantages of the method meant that by the mid-1990s
PCR had begun to replace the original RFLP technique in
many laboratories.
PCR was just finding its forensic legs in February 1993, when
a bomb exploded in the parking garage of the World Trade
Center. The blast killed 6 people and injured more than
1,000. Four days later editors at The New York Times
received a letter claiming responsibility for the attack.
One of the bombers, Mohammed Salameh, was arrested
soon afterward when he went to collect a deposit for the
rental truck that had carried the explosive. Authorities quickly
focused on Salameh’s associate Nidal Ayyad. Included in the
evidence that linked him to the bombing was a trace of saliva
on the envelope flap of the letter received by the Times. The
PCR technique used then was not as discriminating as it
would become later, but experts estimated that only one of
50 people would match the profile found on the envelope,
and Ayyad was one of them. The evidence helped the jury
convict him.
Because the new form of DNA analysis relied on much
shorter repeating segments and therefore lacked the
discriminating power of Jeffreys’s original method,
investigators began to target more STR sites along the
molecule for amplification (thousands are known to exist). To
do so, they simply added primers formulated to attach to
different DNA regions. Two suspects might show the same
pattern for one site, but the chances of a random match
diminished exponentially with each site added. The FBI
eventually settled on the 13 sites that now make up a
standard DNA profile.
The main drawback of PCR was its sensitivity to
contamination. If a few skin cells from a lab worker fell into a
drop of blood that was to be analyzed by the RFLP method,
their DNA would be lost in the much larger quantity of DNA
from the sample itself. But if a contaminant was added to a
tiny trace of evidence intended for PCR analysis, the process
would amplify DNA from the contaminant as well as the
evidence, resulting in a confusing or misleading outcome.
PCR demanded scrupulous evidence handling and
superclean laboratory techniques.
Another difficulty was pinning down exactly how common a
particular pattern might be. DNA analysis never deals in
certainties, only in probabilities. The results indicate that a
particular person’s DNA pattern matches the pattern of an
evidence sample. Statisticians then have to calculate the
chances of a person chosen at random having a similar
match. Is it one in a thousand, one in a million, or one in a
billion?
These numbers are obtained by the product rule. Let’s say
for simplicity that there are 10 possible variants in the
number of repeats at each STR site (the actual number is
usually higher). If the variants are randomly distributed, the
chance of a match at one site is 1 in 10. The chance of
matches at two sites is 1 in 100; at three, 1 in 1,000; at four,
1 in 10,000. Each additional site decreases the likelihood of
a random match by a power of 10. Factoring all 13 STR sites
into the equation, researchers now set the chance of a
random match between two DNA samples at a staggering
one in 575 trillion. When prosecutors explain such numbers
to a jury, DNA evidence becomes very powerful.
In 1994 a case that would make forensic DNA profiling
familiar to millions burst into the headlines. California
prosecutors accused the retired football player Orenthal
James Simpson of stabbing to death his former wife, Nicole
Brown, and her friend Ronald Goldman. DNA became a
central issue in the ensuing trial. The jury sat through what
amounted to a crash course in DNA analysis, and the
technique was mentioned more than 10,000 times in the
transcript.
The analysis produced damning evidence against the
accused. The blood of both victims stained a glove found
outside Simpson’s home. Tests on a stain from Simpson’s
Ford Bronco found that it contained Goldman’s blood. Blood
on a sock found on the floor of Simpson’s bedroom was
identified as Nicole Brown’s. It seemed that the new science
had done an admirable job in unraveling the mystery.
But while the defense did not contest the admissibility of DNA
evidence, Simpson’s “dream team” of lawyers was able to
raise troubling questions about the techniques used to
gather and process the evidence. Much of the testing relied
on the original RFLP technique, which was subject to
laboratory error and interpretation. The newer PCR method,
on the other hand, was acutely susceptible to contamination.
Technicians were notably careless in collecting and handling
samples. One lab worker tested evidence soon after spilling
reference blood drawn from Simpson. These procedural
deficiencies, along with suggestions that Simpson had been
framed by investigators, contributed to the jury’s verdict of
not guilty.
Prosecutors and police realized that scrupulous evidence
handling and rigorous laboratory methods were needed if
DNA evidence was to survive scrutiny. “Based on the O. J.
Simpson case,” the New York City police commissioner
Howard Safir said in 1998, “we had to design a lab and
procedures that are much less subject to challenge.”
In 1996 the federal government began to award grants to
states through its DNA Laboratory Improvement Program.
The funds went for upgrading and standardizing procedures,
and they helped labs switch from RFLP to PCR technology.
That same year the National Research Council gave its
stamp of approval to DNA evidence, declaring that testing
techniques and statistical data had “progressed to the point
where the admissibility of properly collected and analyzed
data should not be in doubt.” Juries were soon routinely
accepting as decisive evidence based on DNA analysis.
In the late 1990s DNA profiling hit the news again, providing
evidence in scandals involving two American Presidents. The
first was almost 200 years old. As early as 1802 Thomas
Jefferson’s enemies had accused him of fathering a child with
his slave Sally Hemings. Proof one way or the other seemed
impossible until 1998. An examination of the DNA of
descendants of Thomas Jefferson’s uncle (Jefferson himself
had no known male heirs) and of that of Hemings’s progeny
suggested that Jefferson had sired at least one of her
children.
Even as it was pointing the finger back into history, DNA
evidence caught in a lie the then current occupant of the
White House, Bill Clinton. Clinton had assured the nation that
he had carried on no sexual relationship with “that woman,”
the White House intern Monica Lewinsky. A DNA profile
ordered by the special prosecutor, Kenneth Starr, proved
otherwise, linking semen stains on Lewinsky’s blue dress to
blood drawn from the President. The irrefutable evidence led
to a shamefaced confession and laid the groundwork for the
first impeachment of an American President in 130 years.
Because of its decisive ability to exclude a suspect, DNA has
always been a powerful tool of exoneration. In 1981 Robert
Clark, 20, was accused of rape, kidnapping, and armed
robbery for a crime that took place in East Atlanta, Georgia.
He proclaimed his innocence, even naming the man from
whom he had bought the victim’s car, which Clark was
assumed to have stolen. But the victim picked him out of a
lineup. He was convicted and sentenced to two life terms plus
20 years.
In 2003 Clark contacted the Innocence Project, an
organization that the lawyers Barry Scheck and Peter Neufeld
had founded in 1992 at New York University’s Benjamin N.
Cardozo School of Law. The nonprofit group provides post-
conviction DNA testing in cases where there is a question of
guilt.
Over the objection of prosecutors, Project lawyers arranged
for DNA analysis in the Clark case. The test proved that
Robert Clark was not the perpetrator of the crime for which
he had been convicted. Instead, it implicated the very man
Clark had named two decades earlier. In December 2005,
after almost 25 years in prison, Clark walked free.
The role of DNA testing in exonerating the wrongfully
convicted has been its most important contribution to justice.
Every mistaken conviction perpetrates a double injustice: an
innocent person locked up, a criminal unpunished. Robert
Clark joined the 175 other defendants who have so far been
cleared by the Innocence Project. This success has spawned
many similar efforts. It has raised questions about the
reliability of other forensic evidence, including confessions
and eyewitness identifications. In 2003 the Illinois governor,
George Ryan, commuted the sentences of all his state’s
death-row inmates after 13 of them were found to be
innocent by DNA testing and other means. Nationwide, DNA
testing alone has cleared more than a dozen persons
wrongfully condemned to death.
Back in 1977, years before DNA profiling was dreamed of, a
robber held up a Richmond, Virginia, business called Shakey’
s Pizza Parlor. The incident turned violent, and the thief
murdered the shop’s owner. In the struggle the assailant left
behind drops of his own blood. With little additional evidence
and no suspects, police placed the case on the list of those
they would probably never solve.
DNA profiling can link a suspect to a crime or exclude him.
But what about those cases where there is no suspect? State
and federal authorities have long kept fingerprint records of
criminals and others so that if a print from an unknown
person is found at a crime scene, they can try to find a
match. They realized that a similar database containing DNA
profiles would allow them to attempt to link evidence from a
crime scene to the profiles of potential suspects.
In 1998 the FBI began a program known as the Combined
DNA Index System, or CODIS, which consolidates DNA
profiles from the states into a national registry. The criteria
for collecting the samples vary by state; many require all
convicted felons and prison inmates to submit a sample. The
PCR technique assigns two figures, indicating the number of
repeats, to each of the 13 DNA sites analyzed. One figure is
a measure of the version of the site inherited from the
mother, the other of that from the father. These 26 numbers
represent a complete, compact profile of a person’s DNA,
allowing for an easily searchable database.
The system has enabled investigators to match crime scene
evidence to the DNA of a potential suspect in more than
30,000 open cases. It has helped solve crimes that are
decades old. By 2006 the FBI had collected 3.2 million
profiles, along with 136,000 evidence samples from unsolved
crimes.
Blood evidence from the Virginia pizza parlor murder was
included in that state’s DNA database in 1996. Then, in
2004, police required a man arrested on a felony gun charge
to provide a blood sample for DNA typing. The two samples
matched. In February 2006 Benjamin Richard Johnson, now
61, was charged with the 29-year-old murder.
A DNA database raises questions of privacy and fairness.
Whose DNA should be included? Only convicted criminals’?
Everyone’s? What should happen to the DNA sample after it
has been analyzed? If a person in a database is a near-
perfect match, should his or her relatives, who might share a
similar profile, become suspects?
Today DNA profiling does not provide any genetic
information about a suspect. Like fingerprinting, it focuses on
a unique but meaningless feature. But the rapid evolution of
genetic science raises the possibility that analysts will soon
be able to infer from DNA physical characteristics like eye
and skin color, approximate height, propensity to illnesses,
and ethnic background. Even a psychological profile is not
out of the question. Is the person prone to anger or violent
action? These possibilities will require society to grapple with
novel ethical questions, but they also hold promise for
effectively using science to bring wrongdoers to justice.
Riding the wave of discoveries in molecular biology and
borrowing the tools developed to map the human genome,
DNA profiling has progressed rapidly over the past decade.
Mechanization has made all aspects of DNA analysis
increasingly automatic. What originally took hours or days
can now be done in minutes. Instead of amplifying and
analyzing a single target section at a time, investigators can
look at all 13 standard sites simultaneously. Instead of
creating bar graphs, they allow a laser detector to profile
segments as the DNA fragments pass down a narrow tube
and let computers analyze the information. Within a few
years, police will be able to analyze evidence and produce a
profile in minutes using hand-held devices right at the scene
of a crime. By tapping into a computerized database, they
may be able to implicate a specific individual almost
immediately.
So powerful is DNA analysis that today a fingerprint may not
only reveal a clue to a perpetrator’s identity in itself but also
yield to investigators a few sloughed skin cells from which
they can extract DNA, develop a profile, and use it to
conclusively nail the culprit.
# # #