Rats! Neuroscientist notches third retraction, this one for using the wrong RNAs

biol psychAmine Bahi, a neuroscience researcher in the United Arab Emirates, has had a third paper retracted.

Here’s the notice for “Blockade of Protein Phosphatase 2B Activity in the Amygdala Increases Anxiety- and Depression-Like Behaviors in Mice,” which was posted on November 19:

This article has been retracted: please see Elsevier Policy on Article Withdrawal (http://www.elsevier.com/locate/withdrawalpolicy).

This article has been retracted at the request of Yale University and the second and third authors, in consultation with Biological Psychiatry Deputy Editor, Eric J. Nestler, MD, PhD.

Reason: The first author informed his co-authors and stated in the article that three short hairpin RNAs (shRNAs) (158–182 bp, 466–490 bp, 1452–1476 bp) had been designed to target the messenger RNA encoding CnA (NM_008915) and had been used in the experiments that formed the basis of the article. The first author has since acknowledged that he used shRNAs directed against rat calcineurin, not mouse calcineurin, in the experiments. Neither the second nor third author participated in or had knowledge of the first author’s actions.

The second an and third authors have re-conducted all experiments reported in the original paper in order to collect new data and republish the findings.

Bahi was the first author of the paper, which was written while he was a postdoc at Yale. The paper has been cited 18 times, according to Thomson Scientific’s Web of Knowledge.

The retraction is Bahi‘s third in three years. In 2011, one of his papers was retracted for “legal issues” that Bahi said were because Novartis didn’t want him to publish his findings. Last year, the cause was lack of animal research committee approval.

15 thoughts on “Rats! Neuroscientist notches third retraction, this one for using the wrong RNAs”

  1. As far as I can tell, NM_008915 is protein phosphatase 3 (Ppp3cc), not calcineurin. From reading the notice it seems that the retraction happened because he used rat instead of mouse calceneurin shRNA. That would not be a big deal at all if he demonstrated that the rat sequence worked since the homology between the two species is very high. It would be a simple correction. Something not right here, unless I am misunderstanding.

      1. shRNA ain’t my thing, but my understanding is that it is delivered at RNA (obviously) and at 3′ sequence. Sequence conservation between species is a lot less than at the protein level. It would only take one base substitution for the whole thing to go pear shaped.

        Happily, it appeared to have worked perfectly! Beautiful downregulation of the target gene and very very miserable mice. And all from an shRNA probe that may well not function at all.

        I would happy wager that this paper and many like it would not replicate at all if the scientists were blinded as to which were the experimental mice and which were the controls.

        BTW, call me Mr Paranoid, but I can’t seem to view pages 142, 143, 144 and 145 of this paper when I try and read it through Science Direct – and I am sure I saw them all a few days ago.

  2. If murine calcineurin mRNA or protein went down in the presence of the shRNAs a simple correction stating that the shRNA sequences were designed to target rat calcineurin should be sufficient… It is odd.

    1. Exactly. Often one is forced to use a off-species siRNAs as they are not always (commercially) available for some species (typically rat suffers more from this than mouse). The homology between rat and mouse genes is also typically very high, so as long as they demonstrate that the siRNA works, it’s no biggie at all. This can’t be the real reason for the retraction. Makes no sense.

      1. They make their own constructs ( I believe) commercial reasons. In terms of RNA high species conservation is not enough – since one degenerate base substitution could prove fatal.

        However, if I have correctly deciphered their methods
        “(158–182 BP; 466–490 BP;
        1452–1476 BP) were designed to target the messenger RNA
        (mRNA) encoding CnA (NM_008915) (31).”
        The 158-182 and the 1452-1476 bp sequences appear identical across species. The 466-490 has an A-G substitution in the middle. I expect that there is enough redundancy built in that it should still work.

        Having said that, I have no faith in these very reductionist molecular approaches to complex behaviours. The field needs to show their methods work when they are applied in a blinded fashion – or replicated by researchers working in a blinded manner – and then what looks so neat and tidy in the journals would rapidly degenerate into something far more messy. In my opinion

        1. A careful look through all of Bahi’s many publications shows them filled with bar graphs and schematic figures. Even when injections of reagents are targeted at neighboring locations in the brain, the sites of injection show up in schematic figures taken from standard atlases. One is left to trust the investigators – we are forced to trust they have accurately presented the locations of these injections; and that trust has been run over, trampled upon and run over again by three retractions. Then, as littlegreyrabbit expresses so well, the issue becomes one of behavioral studies done on animals in which all investigators know who has been manipulated and who has not. Top it off with bar graphs for data (which make Westerns appear as rock-solid when it comes to ease of fabrication) and it all leaves behind the impression this series of studies is one hot mess.

          1. A quick look at Google and Pubmed: in 1.5 years, you got to love the stream of single author papers (almost a dozen!) or with another author who is apparently retired …. Meanwhile, it seems he spent almost 3 years at Yale and 2 in Germany and published 2 papers in total…..

    2. The answer is most likely in your question….. “IF the protein went down”…. I am sure the issue is not just the species problem but data that did not hold after trying to replicate

  3. Look at the blots and you can see there is already a big problem… the bands look upside-down! Yale University + second and third author asking for retraction clearly suggests they have much more concerns than a mix-up of shrna sequences. But as mentioned, this may have been very difficult to prove that the data was not what was found (even by chance) because we trust the investigator. It seems they redid everything so clearly they want to get things right and cleared but Bahi is obviously somebody to read with a grain of salt.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.