[tt] Wired: The Inconvenient Science of Racial DNA Profiling

[Premise Checker]

The Inconvenient Science of Racial DNA Profiling
http://www.wired.com/print/science/discoveries/news/2007/10/dnaprint
[Thanks to Sarah for this.]
7.10.5

By Melba Newsome

On July 16, 2002, a survey crew from the Department of
Transportation found Pam Kinamore's nude, decomposing body in the
area along the banks of the Mississippi known as Whiskey Bay, just
west of Baton Rouge. The police tested the DNA and quickly realized
that they were dealing with a serial killer: the same man who had
killed two other white, middle-class women in the area.

The FBI, Louisiana State Police, Baton Rouge Police Department and
sheriff's departments soon began a massive search. Based on an FBI
profile and a confident eyewitness, the Multi-Agency Homicide Task
Force futilely upended South Louisiana in search of a young white
man who drove a white pick-up truck. They interrogated possible
suspects, knocked on hundreds of doors, held frequent press
conferences and sorted through thousands of tips.

In late December, after a fourth murder, police set up a dragnet to
obtain DNA from some 1200 white men. Authorities spent months and
more than a million dollars running those samples against the
killer's. Still nothing.

In early March, 2003, investigators turned to Tony Frudakis, a
molecular biologist who said he could determine the killer's race
by analyzing his DNA. They were unsure about the science, so,
before giving him the go-ahead, the task force sent Frudakis DNA
swabs taken from 20 people whose race they knew and asked him to
determine their races through blind testing. He nailed every single
one.

Still, when they gathered in the Baton Rouge police department for
a conference call with Frudakis in mid-March, they were not
prepared to hear or accept his conclusions about the killer.

"Your guy has substantial African ancestry," said Frudakis. "He
could be Afro-Caribbean or African American but there is no chance
that this is a Caucasian. No chance at all."

There was a prolonged, stunned silence, followed by a flurry of
questions looking for doubt but Frudakis had none. Would he bet his
life on this, they wanted to know? Absolutely. In fact, he was
certain that the Baton Rouge serial killer was 85 percent
Sub-Saharan African and 15 percent native American.

"This means we're going to turn our investigation in an entirely
different direction," Frudakis recalls someone saying. "Are you
comfortable with that?"

"Yes. I recommend you do that," he said. And now, rather than later
since, in the time it took Frudakis to analyze the sample, the
killer had claimed his fifth victim. The task force followed
Frudakis' advice and, two months later, the killer was in custody.

Colorblind CODIS, Genetic Drift

Tony Frudakis first heard about the Baton Rouge serial killer just
like everyone else outside of Louisiana -- on cable news. As months
went by, the body count climbed, Frudakis followed the case,
thinking "why on earth can't they catch this guy?"

Several years earlier, Frudakis' father was shot when he confronted
a would-be car thief in the driveway of his Long Beach, California,
home. The thief escaped but dropped his driver's license at the
scene and was apprehended quickly. The serial killer had also left
behind his identification in his DNA but, unlike a driver's
license, his genetic ID revealed nothing about his physical
characteristics -- or at least it revealed nothing the police could
use.

The DNA forensic products available at the time could only be used
to match DNA specimens in the CODIS, or Combined DNA Index System,
database which contains about 5 million DNA profiles. If
investigators have a crime scene sample but no suspect, they run it
against those in the database to see if it matches a sample already
on file.

But while CODIS is good at linking the criminals who are already
catalogued from other crimes, the system is useless in identifying
physical characteristics. It says nothing about race. It has been
specifically set up to reveal no racial information whatsoever, in
part so that the test would be consistently accurate irrespective
of race.

But non-scientific considerations also factored into how the system
was established. When the national DNA Advisory Board selected the
gene markers, or DNA sequences which have a known location on a
chromosome, for CODIS, they deliberately chose not to include
markers associated with ancestral geographic origins to avoid any
political maelstrom.

DNAWitness, the test Frudakis applied in the Baton Rouge case, uses
a set of 176 genetic markers selected precisely because they
disclose the most information about physical characteristics. Some
are found primarily in people of African heritage, while others are
found mainly in people of Indo-European, Native American or South
Asian heritage.

No one sequence alone can predict ancestral origin. However, by
looking collectively at hundreds and analyzing the frequency of the
various markers, Frudakis says he could predict genetic ancestry
with 99 percent accuracy.

Based on paleoarcheological evidence and other kinds of DNA
testing, scientists believe we are all derived from populations
that started in Africa and migrated out some 200,000 years ago.
They first settled in the Fertile Crescent, the historic region of
the Middle East flanked by the Mediterranean on the west and the
Euphrates and Tigris rivers on the east.

Various offshoots went in every direction and eventually crossed
the Bearing Strait to America and the populations became sexually
isolated. This process, known as genetic drift, caused markers to
evolve at different frequencies in different populations and gave
rise to the ethnic diversity we see today.

"There is tremendous genetic diversity among other species of
animals but not among humans because our common history is so
recent," he explains. "We're 99.9 percent identical at the level of
our DNA. It's the .1 percent that makes us different and about 1
percent of that .1 percent is different as a function of our
differing history." Frudakis mines that .001 percent to find
distinctive differences that determine genetic ancestry.

Using essentially the same science, DNAPrint helped Oprah Winfrey,
Whoopi Goldberg, Quincy Jones and Chris Tucker trace their lineage
back to Africa for the four-part PBS series, African American
Lives. It's also how, days after the body of 26-year-old Carrie
Lynn Yoder was found at Whiskey Bay, Frudakis was able to conclude
to a statistical certainty that the killer was black.

Racial DNA Profile Leads to Killer

The results from DNAPrint sent the task force scrambling back to
earlier tips about non-white suspects. Three days before Pam
Kinamore's abduction, a black man had tried to rape and murder
Diane Alexander in her home. She survived because her son returned
home and interrupted the attack. Alexander sustained cuts,
fractures and stab wounds but was able to describe the man in
detail. Police never bothered to test the DNA her attacker left
behind. Her case could not possibly be linked to the other murders,
they reasoned, because the suspect was black.

The police had also refused to listen to the pleadings of Collette
Dwyer, who thought she might know the serial killer's identity:
Derrick Todd Lee, a 34 year-old black man with an extensive rap
sheet for domestic violence, assault, stalking and peeping. Lee had
stalked Dwyer for two years after meeting her at the seafood shop
where she worked. One day, he pushed his way into her apartment,
got a drink of water and told her he wanted to "take care" of her.

Lee was arrested after her two children chased him and noticed he
had a gun. He was sent to prison for two years. Dwyer called police
after Pace's murder in May, Kinamore's in July and again in
September following the release of the FBI profile. The police
talked to Lee but didn't bother to take DNA since they were looking
for a white man.

But after the conference call with Frudakis, Lee jumped to the top
of the suspect list. They got a subpoena for his DNA, collected a
cheek swab and a day later, they had their answer: he was their
man. Lee skipped town just ahead of the arrest warrant but was
tracked down in Atlanta and returned to Baton Rouge within days.
"CAUGHT" declared the Baton Rouge Advocate in giant print.

Relatives of the victims described their thrill and relief that a
dangerous killer was finally off the streets, but also frustration
that it has taken so long. Few people knew that the most crucial
piece of evidence was not unearthed by the hapless task force or
forensic scientists but by a drug developer some 800 miles away.

DNAWitness Does Forensics

It takes me a while to find DNAPrint Genomics in spite of its
address on one of the main thoroughfares in Sarasota. The company
is hidden in a small industrial park behind a chain-link fence
across from a busy convenience store. The office suite is marked by
a sign leaning against one wall and a laminated sheet of paper with
the word 'DNAPrint' taped to the glass entry door. The reception
area is sparsely furnished, decorated only with certificates and
plaques.

I am late for my appointment but Frudakis is later. He is out to
lunch and arrives 15 minutes after me. Carrying his lunch bag and
dressed in jeans, hiking shoes and a muted floral shirt, he looks
more like a grad student than a chief scientific officer
responsible for some groundbreaking advancement. While the office
is less than I expect, Frudakis is more -- funny, self-effacing and
candid about his life and work than I expect.

Frudakis earned a PhD in molecular and cell biology from UC
Berkeley in 1995. He spent several years as a research scientist
for Corixa Corporation in Seattle before starting his own company
to develop genomics-based or targeted drugs. The company's first
drug is PT-401, a synthetic version of the hormone produced by the
kidney to promote red-blood-cell production. It can be used to
treat chemotherapy patients and anemia in people with end-stage
renal disease. PT-401 made it through the pre-IND stage, which
comes after animal testing and before Phase I trials, before the
company ran out of money.

With drug approval years away and a gaggle of impatient investors,
Frudakis shifted his focus to forensics in an effort to stay
afloat. The same markers used to infer clinical characteristics
relevant for drug development could also be used to infer phenotype
or physical characteristics that could prove invaluable in
forensics.

By the time he approached the Baton Rouge task force, DNAPrint had
already performed hundreds of dry runs on the test. Its scientists
studied family pedigree to make sure the ancestry traits they were
measuring were indeed passed from one generation to the next. They
conducted population studies, verified the repeatability of the
test, determined the minimum amount of DNA required and completed
more than a 1000 blind trials for various police departments.

When the serial killer's DNA sample arrived at the Sarasota lab,
technicians isolated and amplified the 176 markers, cleaned them up
to remove any primers or other agents, then used the molecular
address to study the sequences at each site. The resulting products
were then deposited into a micro array and scanned by a Beckman
SNPstream. The output was then reviewed and subjected to
quality-control checks. Finally, the scores were calculated and
compiled into a report for the task force.

Since 2003, DNAWitness has been used in more than 150 criminal
cases all across the country and in London. Most remain unresolved.
In several others, however, the science played a crucial role in
narrowing the suspect field and ultimately led to an arrest. Kansas
City, Missouri, police spent four years trying to identify the body
of a 3-year-old black girl. Frudakis determined that the child had
one white grandparent, a clue that ultimately led to the child's
mother, a biracial Oklahoma woman.

When two women were murdered in Napa, California, Frudakis applied
a more advanced version of DNAWitness that uses 1349 genetic
markers to peg the killer as 97 percent Northern European. "The
accuracy of the test was right on," says Napa police commander,
Jeff Tromley. "They described the suspect as a blue-eyed,
blond-haired, white male. When he walked in to the police station,
he was a blue-eyed, blonde-haired white man."

DNA Profiling - Pricey and Dicey

You would think that proven success in solving these types of
horrific crimes would make this technology popular with police,
scientists, defense attorneys and prosecutors alike. But it hasn't.

The most obvious obstacle is price. Cmdr. Tromley, for example, has
a positive opinion of DNAWitness but adds that this does not
necessarily mean his department will use it very often. "This is a
pretty niche product. An in-depth analysis could run from $1500 to
$3000. If you don't need that, then you probably won't go that
far," he says.

Besides the expense, many people who might benefit from DNAWitness
either don't know it exists or are extremely skeptical that it
works. William C. Thompson, Chair of the Department of Criminology,
Law & Society at UC Irvine is a prominent expert on the use of DNA
in criminal trials but was only marginally familiar with this
technology. When I tried to describe how it works, he literally
screamed at me, calling Frudakis a hack and a charlatan who
obviously did not understand statistics.

But even those who believe this can be done are conflicted about
whether it should be done. History is replete with examples of
injustices and inequities that were conscripted into law based on
racial classification. The Civil Rights movement of the 1960's
succeeded in ending legal racial discrimination, in large measure,
by downplaying the significance of race and racial differences. By
the mid-1990s prominent academics and sociologists even went so far
as to say that race did not exist at all.

"Race is a social construct, not a scientific classification," said
an editorial in the May 3, 2001 issue of the New England Journal of
Medicine, adding that "In medicine, there is only one race -- the
human race."

Then, along comes Frudakis with a science that seems to be saying
the opposite.

New York University professor Troy Duster is a member of the
advisory committee on the Ethical, Legal and Social Issues program
at the National Human Genome Research Institute and president of
the American Sociological Association. Duster, who has written
extensively on race and genetics, including the book Back Door to
Eugenics, worries about the proverbial slippery slope.

"Once we start talking about predicting racial background from
genetics, it's not much of a leap to talking about how people
perform based on their DNA -- why they committed that rape or stole
that car or scored higher on that IQ test," says Duster. "In this
society where race is such a powerful idea, once you head down this
path toward predicting race, will the next step be predicting
racial behavior?"

Narrowing the Suspect Field

Frudakis, not surprisingly, takes great pains to explain that those
concerns are overblown. "Our technology is based on the notion that
we all share a common ancestry to Africa from a couple hundred
thousand years ago, that we are all part of the same family tree,"
he says. He also counters critics who say DNAWitness is a high-tech
form of racial profiling. "This is analyzing data derived from a
crime scene. It's a way for police to narrow down their suspect
lists. It isn't used as evidence in trials."

Nevertheless, DNAPrint is still floundering. He says the National
Institutes for Justice denied his grant application because it
believed that this is work that should be left to the government.
It's not clear that the company will be in business a year from
now, or even six months.

"Forensics stinks as a business," Frudakis says bluntly. "Most of
the testing is done by government labs with very little opportunity
for private enterprise. If people valued what we did more, we would
have the funds to expand the databases, learn about more
phenotypes, develop more genetic screens, build more software
systems."

Frudakis still hopes that the company will be able to invest in
more research. RETINOME which predicts iris color with 96 percent
accuracy is on the market and was used very effectively in the Napa
murder case. He has identified the gene sequences associated with
height, and has compiled a database of 5000 digital photographs of
people with almost every racial ancestry combination -- which, one
day, he says could allow him to construct a physical portrait of a
DNA donor, including melanin content, skin color or eye color.

But even the people one might think should be his biggest allies
aren't supporting that, including Tony Clayton, the special
prosecutor who tried one of the Baton Rouge murder cases. Clayton,
who is black, admits that he initially dismissed Frudakis as some
white guy trying to substantiate his racist views. He no longer
believes that and says "had it not been for Frudakis, we would
still be looking for the white guy in the white pick-up truck." But
then he adds, "We've been taught that we're all the same, that we
bleed the same blood. If you subscribe to the (Frudakis) theory,
you're saying we are inherently unequal."

He continues: "If I could push a button and make this technology
disappear, I would."