paris

The Karolinska Institute in Stockholm, Sweden, is distancing itself from the work of one of its former researchers, Ulf Lönn, six months after an independent panel of investigation concluded that he had “most probably” committed scientific fraud.

Lönn denies the charges. But, according to the investigation, about 20 papers published in the past decade by Lönn, who works on gene amplification in breast and bladder cancer, included data that appeared to have been manipulated. “A large number of Lönn's publications must be recalled and disclaimed as unreliable,” says the report.

The Karolinska Institute is now considering introducing a compulsory system of archiving researchers’ raw data with the aim of making manipulation more difficult.

The institute stops short of accusing Lönn formally of misconduct. Hans Wigzell, the vice-chancellor, says that from a legal point of view “we cannot conclude beyond all reasonable doubt that the data could not have come about by means other than fraud.” Nonetheless, Wigzell says, the weight of evidence against the researcher is such that the institute has decided “not to stand by this research, because we consider it is irreproducible science”. He will send letters to funding agencies and scientific journals this week explaining the conclusions of the investigation.

Suspicions were first raised early last year by one of Lönn's colleagues, who felt that the stream of publications emerging from his group exceeded what would be expected from the level of laboratory activity. Shortly afterwards, the institute set up a three-man panel of outside experts, chaired by Jan Pontén, professor of pathology at the University of Uppsala, to investigate the allegations. The panel delivered its report in February.

Lönn believes that he has been legally “cleared”. His lawyer, Frederick Bostrom, from the Stockholm-based firm Lindwall and Nordbage, says that the investigation did not give the scientist “a fair chance” to defend himself, and he may bring a court case.

According to one scientist involved in the investigation, the Karolinska Institute took six months to react to the panel's finding because it was overly concerned about its image. The institute awards Nobel prizes in medicine and physiology. Pontén says that the institute should have alerted scientific journals to the allegations earlier.

Wigzell says the possibility of taking further action was complicated because the institute did not directly employ Lönn, who was funded by the health authorities. Lönn has recently left the institute.