Prompted by PubPeer, biologist corrects three papers

A biologist has corrected three papers that are nearly a decade old, after concerns were raised on PubPeer.

A commenter first posted a comment about an image in one of the papers in 2013; after more comments on other papers appeared in November 2015, author Zoya Avramova at the University of Nebraska-Lincoln received emails alerting her to the threads. She has responded, including to the 2013 comment, noting “the said images should have been assembled more carefully.”

After repeating some of the experiments to verify the results, she has now issued corrections on three papers, about the genetics of model organism Arabidopsis. The papers share a first author, Abdelaty Saleh, who was a postdoc in Avramova’s lab at the time of the work.

The correction notice for “Dynamic and stable histone H3 methylation patterns at the Arabidopsis FLC and AP1 loci,” appearing in the July 2016 volume of Gene, explains: 

A few images bearing the same annotation and similar band patterns were redundantly placed in panels (c) and (e) of Fig. 1. As the original images (2006–2007) are no longer available, the relevant ChIP experiments were repeated by real-time qPCR. The experiments were performed by Dr. Ning Liu, who is credited for the results. The new data providing quantitative estimates of the specific histone modifications at the FLC and AP1 genes in their expressed and non-expressed states support earlier conclusions.

The correction also provides the new data. The 2008 paper has been cited 34 times, according to Thomson Reuters Web of Science.

After the first comment about the paper was posted on PubPeer in November 2015, Avramova responded six days later (she confirmed that the unregistered submission was hers):

All results will be re-confirmed by real time quantitative pCR.

Here’s the correction notice for “The highly similar Arabidopsis homologs of trithorax ATX1 and ATX2 encode proteins with divergent biochemical functions” in Plant Cell, published last year:

Figure 7, panel C (WRKY70 ChIP-PCR assay) has been corrected to clarify that the wt control sample for H3K4m3 was run on a separate gel from the atx1 and atx2 samples, indicated by the space (white line, bottom row). The samples for H3K4me2 (top row) were run together on the same gel. The legend and labeling on the figure have been amended slightly for clarity. The original conclusions of the manuscript are unaltered by these corrections.

Editors note: the corrected figure and accompanying text were reviewed by members of The Plant Cell editorial board.

The paper, published in 2008, has been cited 78 times.

In response to concerns, Avramova wrote on PubPeer in November:

Dr. Saleh was a postdoc in my lab for about a year and a half and as a first author generated most of the results. However, I wrote the papers and as a corresponding author bear the responsibility for all problems.

Going back through the problems, I am inclined to think that the deficiencies pointed out could be explained by lack of sufficient attention when assembling the composite panels. It is possible to mix up images with very similar patterns, and often with the same annotations, when arranging a composite image. I do not imply intention, other than providing a pretty illustration, because I have been following the experiments and scrutinizing the results very closely. Moreover, they have been validated throughout the years by quantitative methods and also by other groups.

However, I did already initiate the process of adding formal corrections, currently with GENE, pending with the others.

Next, the correction notice for “The Arabidopsis homologs of trithorax (ATX1) and enhancer of zeste (CLF) establish ‘bivalent chromatin marks’ at the silent AGAMOUS locus” published in Nucleic Acids Research, explains:

The authors wish to draw the attention to two irregularities in Figures 2a and 4. Both concern errors in duplicating images of ‘empty’ lanes illustrating absence of DNA bands.

The correction provides more details on those irregularities, and notes:

The lower panel of Figure 4 was mistakenly duplicated from Figure 3a, lower left panel. The original gels are no longer available, we therefore wish to remove this panel from the figure and the sentence related to it from the Legend.

The paper has been cited 67 times.

From Avramova’s response to comments on PubPeer:

At the times when PCRs were run in gels, the placement of the samples in lanes did not always follow the order they would later be assembled in text…Although empty lanes carry the same information, I agree that the said images should have been assembled more carefully. The good thing is that with the advent of quantitative real-time PCR such illustrations are not needed anymore eliminating possible mistakes of this type.

Sara Frizzell, the director of Research Compliance at the University of Nebraska Lincoln, told Retraction Watch:

We have received no allegations of research misconduct regarding these papers.

According to first author Saleh’s LinkedIn profile, he has since worked at North Carolina State University then Duke University, and is currently seeking new employment.

Hat tip: Rolf Degen 

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.