Call it bad luck, but the journal Cell has been victimized again by image manipulation. For the second time this month, the publication has retracted a paper whose authors acknowledged that one of them had played around with the figures.
Published in August 2009, the paper, “Population-Level Transcription Cycles Derive from Stochastic Timing of Single-Cell Transcription,” has been cited 16 times, according to Thomson Scientific’s Web of Knowledge. It purported to show using computer modeling that
transcriptional cycling, i.e., periodic assembly of transcription factors and their cofactors and the resulting cyclical accumulation of mRNA, may stem from stochastic timing and sequential activation of transcription in individual cells.
The authors, from several European institutions, presented experimental data to back up their computer model, including multiple figures. Some of those, they now admit, were fabricated — to the point where the whole paper collapsed:
We have recently uncovered instances in several figures where experimental data points were altered or selectively included in or excluded from the analyses; these manipulations do not affect the computational modeling. The first author has acknowledged responsibility for the manipulations. We have now reanalyzed the experimental data and find that they no longer provide statistically significant support for our claim of population-level transcription cycling. We are therefore retracting the paper. We sincerely apologize for any inconvenience this might cause.
The first author of the paper is Tatjana Degenhardt. Degenhardt was affiliated with the University of
Kropio Kuopio, Finland — the home institution of the senior author, Carsten Carlberg — when the study was published but is listed in the retraction notice as now being in the Department of Biological Engineering at the Massachusetts Institute of Technology.
A quick search of Carlberg’s home page revealed another paper he wrote with Degenhardt that was retracted recently, this one in the Journal of Molecular Biology and cited twice. The article, titled “Peroxisome proliferator-activated receptor alpha controls hepatic heme biosynthesis through ALAS1,” was published online in March 2009, and in it the researchers claimed that their data “strongly support a role of PPARalpha in the regulation of human ALAS1 and of five additional genes of the pathway, consequently leading to increased heme synthesis,” heme being an “essential prosthetic group of proteins involved in oxygen transport, energy metabolism and nitric oxide production.”
However, that claim soon began to unravel. Carlberg’s web site in the Department of Biosciences has a section on Degenhardt’s JMB paper that reads like watching a house of cards collapse in slow motion. First, a correction, which takes you to a Word file with ample tracked changes. Then a link to supporting material for the correction, as well as two corrected figures. The last link belongs to the retraction notice, which is scheduled to be published later this month (oddly, there’s no record of the corrected article on Pubmed, and we couldn’t find anything on the JMB site, either), although it evidently became available online in September — which is curious, too, because it’s not in PubMed.
Here is the notice, in full:
In this paper the authors described six of eight genes encoding for human heme biosynthesis pathway as primary targets of the nuclear receptor PPAR α. The most responsive and rate-limiting gene was ALAS1, which has several putative PPAR α binding sites in its regulatory region. It came to the attention of the corresponding author that the first author [Degenhardt] has performed incorrect data handling for the ChIP analysis described in Figs. 4 and 5B and has also used wrong gel pictures as representatives in gel shift assays shown in Fig. 5A. In the following the ChIP data were re-analyzed and the gelshifts were repeated. The claims of Fig. 4 on the binding of PPAR α, RXR α, PGC1 α and pPol II to the transcription start site (TSS) region of the six PPAR α responsive heme biosynthesis pathway still hold true but the numerical values changed. Also the claim of Fig. 5A on in vitro binding of PPAR α -RXR α heterodimers to putative response elements (REs) of the ALAS1 gene could be reproduced. New versions of Figs. 4 and 5 can be found at www.uku.fi/biokem/research/carlberg/Errata/errata.shtml. However, together with the reanalyzed ChIP data of Fig. 5B and the reporter gene data of Fig. 5C, the authors have to change their conclusion about the functional PPAR α binding sites of the ALAS1 gene: not RE1 and RE2, but RE1 and RE3 being located 9 and 0.7 kB upstream of the ALAS1 TSS seem to be the correct sites mediating the effects of PPAR α on the regulation of the gene.
The chief editor and authors requested that this paper be retracted and sincerely apologize for these errors and for any inconvenience this might cause.
Degenhardt evidently left Finland for the United States, where she seems to have been a postdoc in the lab of Ernest Fraenkel, but you wouldn’t know it from Fraenkel’s web site. The only trace of her is in his CV, a copy of which is available online, where she is listed as having started in 2009 and working through “present.”
We have tried to contact Carlberg and will update this post if we reach him.
We emailed Degenhardt to see if she had anything to say about the retractions. She replied, saying she no longer works at MIT, but declined to answer any questions about the paper.
[T]he process is not yet dealt with entirely, I would rather not comment on it. I am very sorry that it happened, but is also very difficult to explain.
We spoke by phone with Fraenkel and he confirmed that Degenhardt had been with his lab until the summer, but he would not comment on the reasons for her leaving. Fraenkel said he was “absolutely not” worried about any problems with research she conducted with him while at MIT, although he admitted that the two had not worked together on any manuscripts.
Meanwhile, we reached out to Cell editor Emilie Marcus (no relation to AM, we’ve established) to talk about the fact that this is the second time within a month that the journal retracted a paper by a team that accused the first author of falsifying dozens of figures.
We asked Marcus whether the two recent retractions might prompt some kind of change at the journal regarding its policies for images and figures. We will update this post if/when she responds.
Hat tip to @biochembelle for alerting us to the retraction in Cell, which joins our list of journals we wish would send out a press release when they retract studies. Please seen an update with comments from the PI of the Cell paper.