Checking stem-cell work poses challenge

It's a pillar of the scientific method: Fraud
or errors in a scientific finding from one lab are exposed when other
labs can't confirm the work. But that process has been hobbled in the
stem-cell research scandal in South Korea, some scientists say. The
reasons include the technical challenges of the work and restrictions
by the U.S. government. 

Researcher Hwang Woo-suk of Seoul National
University resigned his post Friday after investigators concluded he
had faked results in a headline-grabbing report published in May.

The original study said Hwang's lab had created
11 lines of embryonic stem cells genetically matched to individual
patients. The investigators concluded that at least nine of those lines
were bogus.

The fraud findings cast doubt on Hwang's other
work, including his claim in February 2004 to have been the first to
clone human embryos and extract stem cells from them.

Nobody has yet published a success in
duplicating or extending that experiment. Scientists say such follow-up
work is key to rooting out errors and falsehoods in scientific
findings. If nobody can duplicate or extend the results of an
experiment, then questions about the original findings arise.

But the failure to confirm the 2004 Hwang
research doesn't necessarily mean it was bogus, cautions Dr. Curt Civin
of Johns Hopkins University Medical School, editor of the journal Stem Cells. He pointed to several obstacles to trying to do such followup work.

Such attemps are off-limits in many U.S. labs
because the government restricts federal money for human embryonic stem
cell experiments, he said. Labs that depend on federal money cannot use
it to create new embryonic cell lines as Hwang claims he did.

The federal policy dampens the possibilities for
followup work in other countries because "a lot of countries look to us
for guidelines on what they should do," Civin said.

What's more, few labs are prepared technically to do such work, he said.

"If it were something that was important and
could be done by many labs, and was not prohibited," he said, many
scientists would have attempted it by now.

John Gearhart, also a Hopkins researcher, said
another impediment is that Hwang's lab has not shared enough details of
its technology to allow other labs to duplicate it. And apart from
technical acumen, few labs in the world have permission to do work that
would build on the Korean report, he said.

Maybe some labs have tried and failed but kept
quiet about it, Civin said, because the scientists "didn't feel like
saying they weren't as good as the South Koreans."

Indeed, the technical expertise of Hwang's lab
has often been mentioned as a crucial ingredient in its reported
success. Hwang has attributed his lab workers' dexterity to lots of
practice with steel chopsticks.

Visitors to Hwang's lab report "they have
technical competence and abilities far beyond what most American labs
have, and I think that's true," said Thomas Cech, a Nobel Prize winner
who heads the Howard Hughes Medical Institute. "People have actually
watched them manipulate these cells ... and say they're really
extraordinarily dexterous and skilled technically."

So if other scientists can't get such
experiments to work, he said, they may conclude "I'm not surprised
because I saw these people from Korea and they're so much better at
manipulating cells than I am."

The journal Science, which had published
Hwang's cloned-embryo study, said last week that it was investigating
the work. At issue are two photographs that appear identical to images
that another journal published in a previous study on another topic.

Earlier, a co-author of the now-discredited Hwang paper on the 11 cell lines had accused Hwang of fraud.

Duplicated images and tips from labs have led to the exposure of scientific fraud in the past.

In one of the most spectacular cases in recent
years, Bell Labs researcher Jan Hendrik Schon was fired in 2002 after a
bunch of reported experiments produced results that looked eerily
similar when graphs from his papers were compared.

And an accusation of scientific misconduct from
a research assistant launched an investigation in 2000 of a professor
at the University of Vermont College of Medicine. After resigning in
2001, Eric T. Poehlman pleaded guilty in April of this year to
fabricating research data, including on such topics as menopause, aging
and hormone supplements.