Retraction Watch

Tracking retractions as a window into the scientific process

A journal printed a sharp critique of a paper it had published. If only it had checked with the authors first.

with 7 comments

In 2016, researchers published a paper showing that an RNA molecule may be overactive in breast tumor tissue. But after reading the paper, three biologists believed the data supported the opposite conclusion.

What happened after that is a tale of misunderstandings and unnecessarily bruised feelings. We’ve seen plenty of cases where researchers ignore criticism, which at first glance seemed to be the case here. But upon closer inspection, it wasn’t.

The concerned readers—Debomoy Lahiri, Kumar Sambamurti and Bryan Maloney—contacted the authors in June 2017 for clarification on the apparent contradiction between the results and data. When they didn’t hear back, the trio submitted a letter to the journal, Gene, a week later, highlighting potential “critical errors” in the article. According to the journal editor, in the interest of transparency, it published the letter—which contains highly pointed critiques of the paper—in September. But the letter appears without comment or clarification from the paper’s authors.

The first author says the journal didn’t reach out before publishing the letter; what’s more, the first and the corresponding authors argue the critics simply misinterpreted their findings, and the first author has asked them to withdraw the letter. The critics say they were confused by misleading legends on some figures. The journal editor says he wishes the entire dialogue had stayed private.

Here’s what sparked confusion over the paper, “MiR-346 promotes the biological function of breast cancer cells by targeting SRCIN1 and reduces chemosensitivity to docetaxel,” which has been cited three times, according to Clarivate Analytics’ Web of Science. First author Fan Yang and corresponding author Jin-hai Tang—who both work at Nanjing Medical University Affiliated Cancer Hospital in China—reported that MiR-346 may fuel the spread of breast cancer. But Lahiri, Sambamurti and Maloney believed two figures in the paper showed that MiR-346 decreases, not increases, in cancer tissue. As they write in the letter, “When figures and data contradict text: MiR346 is apparently reduced in breast cancer tissue, contrary to claims by a paper’s author:”

The problem with these statements is that they contradict the actual data presented in the paper!…In its current state, the text will certainly create confusion in the field and lead to incorrect assumptions…

Yang and Tang told us the critics simply misinterpreted the data. The explanation lies in the statistical tools they used: The authors say they had calculated the expression of miR-346 in tumor tissue and healthy tissue using the widely accepted method, 2−ΔΔCt, but had depicted this activity in the two figures in question using a different method, ΔCt. Tang and Yang explained they used the ΔCt method “for the beauty and saliency of the figure,” but that the 2−ΔΔCt method reveals that “the expression of miR-346 in tumor tissue is about two times higher than in normal tissue.”

Yang also explained why the critics didn’t get a response when they first emailed in June:

… for personal reasons, I haven’t been here and read the mail for some time, so I didn’t find out they had contacted me before.

Yang also said the journal “didn’t contact me directly.”

One misunderstanding after another

The editor-in-chief of Gene, Andre van Wijnen, told us he couldn’t remember if Tang and Yang reviewed the letter before it was published:

By procedure, our intent is that authors have a chance at a reply, and typically this is done by inviting the authors as ‘reviewers’ by routing the letter as if it is a full paper submission. The original authors may or may not have responded in time, or they may have responded and provided non-constructive comments.

Regardless, van Wijnen explained, “we proceeded with the Letter for the reason that it exposed major weaknesses:”

Personally, I think that the authors should have just kept this a private conversation between colleagues and then reach some kind of consensus followed by a corrigendum that clarifies uncertainties. At least that is how I think science should be conducted.

The idea behind the corresponding author is that this is the point person who will handle any inquiries to the work. Why anyone would want to escalate disagreements about data to the level of a public letter that is somewhat scathing is beyond me, but since the journal offers the format, it would be a form of censoring if letter writers did not have a forum in GENE to express their view.

After we reached out, Yang emailed Lahiri October 4 to address the concerns with the 2016 paper, and to ask why Lahiri hadn’t contacted the authors directly. Yang also asked that Lahiri and colleagues to withdraw their letter.

Lahiri responded to Yang on October 7:

We are grateful for providing us with an explanation for the apparent misinterpretation of your study. We did not contact the editors of Gene frivolously or as any sort of disrespect.

Lahiri told us:

We agree that we missed the fact that they presented ΔCt for one half of the figures while using 2-ΔΔCt, for the remainder. In our defense, all we have is one small representation (ΔCt) embedded within a large label on the Y axis, and no explanation in the figure legend. Figure legends do not indicate that authors were changing the presentation of these values for “saliency and beauty”. Without the clear notation of the exception in the figure legend, the readers are forced to depend on a tiny change in the Y axis of the figure and depend on minutia of the methodology. Moreover, “Materials and Methods” of their paper just lists one method of analysis: 2−ΔΔCt, and not ΔCt.

Lahiri also told us:

we are grateful for the opportunity to clarify events and bring to light a potentially severe shortcoming in the published article. We feel that it is an obligation for authors to provide clarification in the form of revised figures and legends for the benefit of future readers.

Regarding the fate of the paper and letter, van Wijnen told us:

This Letter is there on its own merit, and readers can make up their mind whether the accepted paper should stand. If the arguments in the Letter are true, you would expect the authors to volunteer and request a retraction.  If they do not, then the Letter is there for all to see, and the work is critically debunked. It is not about necessarily about publishing, but whether a paper is useful and citable.

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 daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

Written by Victoria Stern

October 11th, 2017 at 8:05 am

Comments
  • Paul A Thompson October 11, 2017 at 8:17 am

    If you calculate and test using one method, and use another method to “illustrate” what you calculate, you are engaging in a misrepresentation of one sort. In science, using a figure which more “beauty and saliency” is not appropriate. The figure should be corrected to reflect the actual calculated value.

  • TL October 11, 2017 at 8:31 am

    “The trio submitted a letter to the journal, Gene, a week later”

    Is this a joke? Surely you should allow the authors more than one week to respond?

    I wonder if the corresponding authors’ email addresses at foxmail [dot] com and 163 [dot] com (common Chinese free email providers) had anything to do with the apparent failure of communication?

  • Neuroskeptic October 11, 2017 at 9:26 am

    The authors are responsible for this, as far as I can see.

    “For personal reasons, I haven’t been here and read the mail for some time, so I didn’t find out they had contacted me before.”

    Lord knows, there are more important things in our personal lives than research, but still, if you don’t respond to a critical comment, don’t be surprised if you get criticized without being able to respond.

  • David B. Karpf, MD October 11, 2017 at 12:09 pm

    As far as I am concerned, this retractions should be reserved for meaningful scientific misconduct. I might have expected other labs working in the same field to have submitted a letter pointing out the potential for misunderstanding the difference in the y-axis of the 2 figures. What I DO NOT understand, and do not support, is the publication of an extremely critical letter, that basically states that the conclusions in the published manuscript, are completely wrong without giving the authors of the paper a chance to respond. This result did not involve patient treatment – there was no urgency to publish this severely critical letter one week after publication of the paper. This is social media run amok – I think Gene should retract the letter, and the authors should submit a revised figure legend more clearly explaining the difference in the axes, and probably adding a 3rd figure of the “primary” mrna expression result (delta delta).

    • Michael October 11, 2017 at 1:11 pm

      To clarify, the letter was published one full year after the paper was available online, not a week later. Letters to the Editor has their place in the pre-internet era, now they are an anachronism.

      • Elf October 11, 2017 at 3:07 pm

        It says the “concerned readers” contacted the authors in June 2017, and submitted the letter to the journal one week later. The publication history for the letter confirms it was submitted on 14 June 2017:
        “Received 14 June 2017, Accepted 20 August 2017, Available online 6 September 2017”

        More worrying to me is the editor’s claim that he “couldn’t remember “ if the original authors reviewed the letter during the peer review process (ie between 14 June & 20 August 2017, based on publication history. Isn’t this information captured in a manuscript management system?

  • Michael October 11, 2017 at 12:43 pm

    This episode vividly demonstrates the need for a widespread adoption of transparent and immediate post-publication peer review, which includes a system to acknowledge/reward those who take the time to thoroughly read a study and make useful comments (positive or negative). Yes, it will take time and some false starts to find a system to “review the reviews” but the scientific community needs to get serious about considering published work to be just the start of the process, not the end. PubPeer has done important work, but remains a pariah for many. The concept of PPPR needs to find more widespread acceptance in the community.

  • Post a comment

    Threaded commenting powered by interconnect/it code.