Balter, Michael. "Data in Key Papers Cannot Be Reproduced," Science 283 (26 March 1999),


Balter, Michael. "Data in Key Papers Cannot Be Reproduced," Science 283 (26 March 1999),
pp. 1987, 1988.

New findings, published last week, appear to confirm suspicions that several key papers in a hot
area of plant development were fatally compromised by scientific fraud. The results, published in
the March issue of Plant Journal, stem from an investigation at the Max Planck Institute for Plant
Breeding Research in Cologne, Germany, which concluded last year that a laboratory technician
falsified experiments forming the basis of 10 publications going back to 1992. The technician,
Inge Czaja, and the leader of the group in which she worked, Richard Walden, resigned in early
1998 in the wake of the scandal, although Walden has never been accused of participating in the

In the Plant Journal article, a team of researchers at the Cologne institute, along with colleagues
from other European labs, report on their attempts to repeat key experiments in eight papers
published in Science, EMBO Journal, the Proceedings of the National Academy of Sciences
(PEAS), Trends in Plant Science, and Plant Journal. The authors could not reproduce the most
central findings. Two other papers from the institute, which had originally appeared in Nature
and PEAS in 1997, were retracted last year by most of their authors after their findings also could
not be reproduced.

"I can no longer believe any parts of the of the data in any parts of the papers," says plant
biologist Alan Jones of the University of North Carolina, Chapel Hill, who adds that the new
findings will have "a negative effect on the field'" because "major conclusions were drawn" from
the papers. The lead author of the Plant Journal report, plant researcher Jeff Schell - who is head
of the department in which the Walden group worked and a co-author on the disputed papers -
agrees that all the major findings were "subject to falsification." Nevertheless, Jones, Schell, and
other researchers stress that the basic techniques used in the research - which were pioneered by
Walden and other colleagues - remain valid and are being enthusiastically used by other
researchers. "This technology has been very influential," says plant molecular geneticist George
Coupland of the John Innes Centre in Norwich, United Kingdom.

The affair dates from the early 1990s, when Walden and his co-workers pioneered a new way to
study the actions of plant genes. The technique, called activation T-DNA tagging, creates
mutations by inserting DNA from the soil bacterium Agrobacterium tumefaciens, which induces
plant tumors, into the genome of plants they wish to study. They found that genes flanking this
inserted foreign DNA were "overexpressed"; that is, they produced much higher levels of
proteins than normal, allowing those and their protein products to be studied much more easily

With this method, Walden and his co-workers began trying to decipher the poorly understood
mechanisms of action of twolent hormones auxin and cytokinin that control plant cell division
and grown To do this the team produced numerous mutants of tobacco plants which they thought
were capable of growing independently of the presence of these two hormones. Using these
mutants, the team isolated a number of genes, proteins, and other factors that appeared to
stimulate plant growth "downstream" of the hormones - and thus were implicated in the
hormones' mechanism of action.
It now appears, however, that these mutants were not capable of independent growth after all.
The investigation carried out at the institute concluded that Czaja added plant growth factors to
culture media used in the experiments and manipulated the experiments to make it appear that
cultured plant cells were capable of auxin- and cytokin-inindependent cell division. (Czaja, who
was also a co-author on the papers, declined to comment when contacted by Science.) Serious
suspicions had been raised by earb 1998, when researchers at the institute were unable to repeat
results stemming from the technicians work Walden and his co-workers began investigating and
soon concluded that at least some of the results had been faked.

In March 1998, Walden informally let other plant researchers know that there were potential
problems with the work, and the following month he, Schell, and another co-worker published an
initial warning about the data in Trends in Plant Science. Nevertheless, under strict new rules on
scientific misconduct adopted by the Max Planck Society in November 1997, institute officials
sought, and received, Walden's resignation. "There were ample signs that [Walden] did not
exercise proper responsibility for his group," says Heinz Saedler, a co-director of the Cologne
institute. (Walden, who now works at a research institute in the United Kingdom, told Science he
preferred not to comment on the affair.)

Despite the dramatic findings in this month's Plant Journal report, Schell says the group has no
immediate plans to publish retractions of the eight papers in the journals in which they originally
appeared. "This article is about the only thing we were planning to do. The main thing is to get
our science going again." On the other hand, Schell adds, if the journals themselves asked for
retractions, "I would consider it very seriously." But some editors of the journals involved say
they believe the co-authors should submit letters stating that the results could not be reproduced.
John Tooze, co-executive editor of EMBO Journal, says that although the journal has no
hard-and-fast policy about retractions, it would be "common sense" for the authors to contact the
journals irn~olved. "A statement in each of the journals from the authors would be an appropriate
thing to do," he says. And Floyd Bloom, editor-in-chief of Science where three of the eight
papers appeared says that "we would have expected Dr. Schell or his institution to contact us
when the results that had been published in Science were conclusively identified as suspect. We
will be discussing the possible need for retractions of the papers that Dr. Schell and his
collaborators published in Science with him, and will act accordingly."

Jones says that, in retrospect, flaws in some of these papers might have been spotted with closer
review. For example, in the Plant Journal study the researchers used a second assay technique -
incorporation of the DNA building block thymidine into plant cells - in addition to a
cell-counting method used in the original work to determine whether cell division had occurred.
"In hindsight, why wasn't the thymidine incorporation done originally; why didn't the reviewers
call for that9" Jones asks. On the other hand, he says, ‘hindsight isn't fair.... When the papers
came out I was extremely enthusiastic."