U Maryland virus researcher up to 13 retractions

A veterinary researcher at the University of Maryland has lost seven papers for problematic images and other issues, bringing his retraction total to 13. 

Siba Samal, who studies viruses and vaccines, lost four of his articles in March after journals determined that figures in the papers were unreliable. And he was a co-author on papers flagged in a U.S. Office of Research Integrity finding of misconduct by another researcher.

Some of those questions were raised — first, it seems — by Elisabeth Bik, who in November 2015 reported her concerns to PLOS ONE about a different one of Samal’s articles, which she also flagged on PubPeer. 

The journal is now retracting that article, titled “Evaluation of the replication, pathogenicity, and immunogenicity of Avian Paramyxovirus (APMV) serotypes 2, 3, 4, 5, 7, and 9 in Rhesus macaques.” According to the retraction notice: 

After this article [1] was published, concerns were raised about results reported in Fig 1.

Specifically:

The following lanes appear similar:

Day 0 lanes for Animals #2, 3 in the rAPMV-4 panels

Day 0 data reported for Animal #1, rAPMV-4, appear to match the image data for Day 0, Animal #4, rAPMV-4/Fc-BC. This issue was confirmed in the University of Maryland’s investigation of this work.

Day 0 lanes for Animals #1, 2, 3, 4 in WtAPMV-7 panel and Animals #1, 2, 3, 4 in the rAPMV-7/Fcs-5B panel

Day 28 data for Animals #1 and #3 in the WtAPMV-7 panel

There appear to be vertical discontinuities suggestive of image splicing between lanes in all panels of Fig 1.

Concerns were raised about controls and aspects of the experimental design for the Fig 1 results. No loading control data or pre-immune/no antibody control blots were reported to support the western blot results in Fig 1. Furthermore, the authors explained that virus samples for each of the APMV strains were run together on the same blot; the membranes were cut into individual strips and incubated individually with Day 0 or Day 28 sera from each monkey; after blot development by chemiluminescence, figures were generated by compiling data obtained on multiple blot strips. This image splicing was not declared in the figure legend, and the validity of comparisons between Day 0 and Day 28 results is in question since results were obtained using different blot strips and no loading control data were reported.

The authors commented that Day 0 data looked similar across experiments and that the same control data were reused to represent different results due to errors in figure assembly. Regarding the Day 28 results (Animals #1 and #3 in the WtAPMV-7), the authors stated that the original results for these animals appear similar but they stand by these results as reported in the figure.

The University of Maryland investigated this work and confirmed that several lanes are duplicated in this figure, including control (Day 0) lanes in several panels and the two Day 28 lanes in the WtAPMV-7 panel.

Raw image data were provided in support of Fig 1. The image data supported the Day 28 results reported in most panels of the figure. The Day 0 data did not match the results in the figure in most cases, and for some control lanes we were unable to confirm verifiable image content. Overall, the issues about experimental controls and Day 28 results for WtAPMV-7 were not resolved, and we remain concerned about the extent of image data reuse in the published figure.

In light of the above concerns, the PLOS ONE Editors retract this article.

UJB agreed with retraction. SKK, SKS, and PLC did not agree with retraction and stand by the article’s findings. PLC apologizes for the issues with the article. The other authors either could not be reached or did not respond directly.

PLOS ONE also is retracting the following six papers by Samal’s group, which, as Bik notes, involve 

a range of problems, varying from reuse of images across different papers (representing different experiments), to splices in blot, animal ethics, and duplications within blots (which are true manipulations)

They are:  

• Experimental infection of mice with avian paramyxovirus serotypes 1 to 9 
• Replication, Neurotropism, and Pathogenicity of Avian Paramyxovirus Serotypes 1-9 in Chickens and Ducks
• Newcastle disease virus fusion protein is the major contributor to protective immunity of genotype-matched vaccine
• Mutations in the Fusion Protein Cleavage Site of Avian Paramyxovirus Serotype 4 Confer Increased Replication and Syncytium Formation In Vitro but Not Increased Replication and Pathogenicity in Chickens and Ducks
• Effects of naturally occurring six- and twelve-nucleotide inserts on newcastle disease virus replication and pathogenesis
• Comparative immunogenicity of HIV-1 gp160, gp140 and gp120 expressed by live attenuated newcastle disease virus vector

Four of the newly retracted articles include co-author Shin-Hee Kim, a former scientist at Maryland who in May was found guilty by the U.S. Office of Research Integrity of having committed misconduct in seven papers and two federal grants. 

Samal tells us that ORI cleared him of misconduct:

I strongly believe that any intentional manipulation of data in a research paper should be dealt with firmly and the person(s) involved should be punished accordingly. But if the mistake happened due to a misunderstanding or unintentionally, then they should be corrected but not retracted. In those papers where I disagree with the publisher’s decision to retract, because the data were reproducible both by my lab and another lab and I’m confident about the data. Moreover, the mistakes did not change any conclusions of the papers.  Whereas the papers I have not tried to reproduce because of lack of reagents and having even slight doubt on the data, I fully agreed with the retraction.

Regarding ORI’s findings, I have received their final report with the conclusion that they did not find any research misconduct on my part. This finding is in complete agreement with the university’s final findings.  In both investigations, they found all the mistakes were beyond human capacity to be detected by the PI.

Dr. Kim was a research assistant professor in my lab while all those papers were produced.  As a senior research scientist, I trusted her data and never had any suspicion on her integrity during her almost 10 years time in my lab.  unfortunately, she is the one who committed the majority of the research misconduct and therefore the problems are similar.

Samal laid the remainder of the misconduct at the feet of another co-author at Maryland, Sunil Khattar. Khattar did not return a request for comment. 

As Bik has found out: 

All the Shin-Hee Kim papers with PubPeer posts also have Siba K Samal as the last author, while not all Siba K Samal papers with PubPeer posts (or with retractions) have SHK as an author. So it appears that SHK is not the only person who might have committed fraud. 

Samal and Kim also have an expression of concern in PLOS ONE, for this 2017 paper titled “Evaluation of fusion protein cleavage site sequences of Newcastle disease virus in genotype matched vaccines”: 

After this article [1] was published, it came to light that the Data Availability Statement is inaccurate: the underlying data are not included with the published article. Additionally, there are six statements in the Results that rely on data not shown which is not allowable per the journal’s Data Availability Policy.

The authors clarified that the underlying data for this study are not available, and so the article does not comply with PLOS ONE’s Data Availability Policy.The PLOS ONE Editors issue this Expression of Concern to notify readers of this issue. We regret that this was not addressed prior to publication.

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].