Å®ÉúСÊÓƵ

Sloppy laboratory work creates the greatest risk of error when samples
of DNA are used as evidence in court, according to a new report from the
National Academy of Sciences. The academy calls DNA analysis a ‘powerful
tool’, but it also calls for caution in interpreting DNA results and regulation
of laboratories that carry out analysis of DNA.

Supporters and opponents of DNA ‘fingerprinting’ in the US have been
fighting an increasingly bitter battle in courtrooms, scientific journals
and the media. Victor McKusick, of Johns Hopkins University in Baltimore,
who chaired the NAS panel, says he hopes the study ‘puts the controversy
to rest’, but that seems unlikely.

The report recommends regulation and testing of all examiners and laboratories
that produce evidence for use in court. Outside inspectors should conduct
this ‘blind proficiency testing’, and the results should be made public.

The FBI should change the way it calculates the probability that a random
person’s DNA would match samples found at the scene of a crime, says the
report. The new procedures would increase the estimated probability of a
chance match from 1 in many millions to 1 in perhaps 10 000.

The committee called for detailed study of 15 to 20 different ethnic
groups in the US. Two people from the same ethnic group are more likely
to have a close match, yet in court cases, the likelihood of a random match
between DNA from two different people is usually estimated using data from
the general population.

Last month, a new analysis of genetic data from an isolated indigenous
group in South America produced a dramatic illustration of this point. Kenneth
Kidd of Yale University collected genetic profiles of members of a tribe
called the Karitiana. Kidd has reported in courtroom testimony and in an
article in the journal Science that none of the 54 people in this population
had identical genetic profiles. A strong supporter of using DNA in the courts,
Kidd used this data to show how rare random matches are, even among relatives,
and thus how reliable DNA techniques are.

But Kidd’s data contained a few surprises for Laurence Mueller, an evolutionary
biologist at the University of California, Irvine. Mueller assumed that
Kidd had looked for a match at all seven locations, or loci, on the chromosome
where he had collected data. Mueller thought there might be some matches
if only three or four loci were considered, as is typical in criminal cases.
He programmed his computer to compare the 1431 possible combinations of
pairs in this population, using the FBI’s criteria for a match.

When the results emerged, Mueller was astonished to find 322 pairs that
matched at four loci, 61 matching pairs at five loci, five pairs of Indians
that matched at six loci, and two pairs that matched at all seven loci.
This was exactly what Kidd had testified did not occur. Kidd was not available
for comment, but according to court papers filed in an Ohio case, does not
dispute Mueller’s analysis.

By contrast, the academy’s report says that a search of the FBI’s database,
which includes more than 7 million genetic profiles, produced only one pair
that showed an ‘exact match’ at three loci.

Mueller’s analysis has become one of the prime exhibits in an attempt
to set aside the verdict in a 1989 murder case in Ohio where DNA evidence
was used. Kidd testified for the prosecution, and two prominent population
geneticists, Richard Lewontin of Harvard and Daniel Hartl of Washington
University in St Louis, testified for thedefence.

The case has become a focus for advocates and opponents of DNA evidence.
The defence argues that prosecutors hid the fact that several of its expert
witnesses, including Thomas Caskey of the Baylor College of Medicine, had
applied for grants from the Justice Department to work on DNA forensic techniques
at the time of the trial. This fact, says the defence, could have discredited
their testimony. The grantswere later approved.

Caskey was a member of the National Academy’s study of DNA fingerprinting
until a few months ago. He resigned because of worries that his ties to
several companies that are involved in DNA analysis might pose a conflict
of interest.

Defence lawyers also say that the FBI, the Justice Department and its
hired scientists have conspired to intimidate critical scientists and suppress
their work. In a supporting affidavit, Mueller says that a prosecutor from
Oakland, on learning that Mueller had submitted papers to the journals Science
and Genetics, sent letters to both journals with ‘disparaging comments’
about him.

Hartl, in another affidavit, describes how James Wooley, prosecutor
in the Ohio case, phoned and tried to get him to withdraw a paper that was
submitted to Science. Wooley had received a copy of the paper from an FBI
agent who obtained it at a trial where it was entered into the record. Hartl
says he found the conversation ‘intimidating’. Caskey and Kidd, according
to a news account in Science, later ‘cornered’ one of the journal’s editors
at a conference and urged that Science should not publish the paper without
a rebuttal. Daniel Koshland, the journal’s editor, commissioned a rebuttal
article by Kidd and Ranajit Chakraborty of the University of Texas.

The final piece of evidence the defence puts forward is the FBI’s attempt
to influence the academy’s report. John Hicks, assistant director of the
FBI, wrote last October that some committee members gave him parts of the
draft report because they thought the draft ‘lacked objectivity and balance’.
Hicks, who is not a geneticist, then sent the committee an eight-page letter
criticising the draft and urging the committee to ‘include scientific viewpoints
that had been omitted’.

Prosecutors call the Ohio defence briefs ‘an all-out, bare knuckles,
take-no prisoners assault’ by the ‘anti-DNA lobby’ that would merit a libel
suit if they were not attached to a court motion. He says thedebate over
the use of DNA forensictechniques ‘isn’t a scientific debate atall’. He
sees it as a battle between scientists who are ‘pro-defence’ and others
who are ‘pro-prosecution’.