Correction

In July 2002, an article entitled "Expression of Oct-6, a POU III Domain Transcription Factor, in Schizophrenia" was published in the Journal (159:1174–1182). We wish to retract one of the conclusions.

Since publishing this study, our laboratory at the Institute of Psychiatry, King"s College London has had difficulty reproducing some of the findings. In particular, the primary observation that Oct-6 is ectopically expressed in schizophrenic brain tissue could not be reproduced in the Stanley series of postmortem samples. These negative findings were recently published (1) .

This failure to reproduce led us to re-examine the data underlying the original publication. We now conclude as follows:

Figure 1 in the original paper presented the characterization of an anti-Oct-6 antibody by electromobility shift assay. These data were contributed by one of the co-authors (D.M.) from his laboratory and are completely correct as far as we are aware.

Figure 2, the principal component of the study, illustrated the immunohistochemical analysis of schizophrenic and control tissue and was conducted in the King"s College Laboratory. These data have proved irreproducible in other schizophrenia samples (1), but we have no specific evidence that they are incorrect. Material from the original samples is no longer available, so unfortunately a direct rerun of these precise experiments is not possible. Nonetheless, we would say that these data should be regarded as unreliable.

Figure 3 presented an immunoblot analysis of Oct-6 expression in schizophrenic and control tissue. We have prima facie evidence that these data are fraudulent. There are two reasons for reaching this conclusion. First, close examination of the lanes on this figure indicate that they have been manipulated and cannot be what they purport. Second, we have been able to track the derivation of this figure from the primary data, and that analysis reveals that the data have been manipulated. Needless to say, these data should not be considered reliable.

In light of these revelations, we retract the finding that Oct-6 is dysregulated in schizophrenic brain tissue.

The appropriate authorities at King"s College were informed of the suspicion of fraud, and an investigation was carried out under the College"s "Regulations for Investigating and Resolving Allegations of Research Misconduct." That investigation, assisted by the production of a report from an independent expert, concluded unequivocally that some primary data produced in the King"s College Laboratory by Dr. Maria Ilia had either been falsified or had been wrongfully manipulated to produce a misleading analysis. As a result of this, the signatories to this letter now formally retract the paper.

We would like to make clear that the source of the data was a single researcher in the King"s College Laboratory. The investigation by the College has attached no suspicion at all to the other authors of the paper. That includes those currently and previously at the Institute of Psychiatry or in Rotterdam.

Finally, we wish to apologize sincerely to The American Journal of Psychiatry and its readership. You have a right to expect the highest standards of academic practice from authors whose work is submitted to you. Clearly, there has been a substantial failure in this regard, and we are embarrassed and distressed that this has occurred.

The Journal sought comment from Dr. Ilia on this matter, but none was received.