A reader informed the journal that there were several potential issues with the blots in this paper, which were also highlighted on the PubPeer website. Possible band duplications were highlighted in: p21 and actin blots in OSE457 and MCF7 panels in Fig. 1A; actin blots in A2780 and OVCAR5 panels in Fig. 1B; actin blots in IOSE 166 h and WI38 panels in Fig. 2A; T-AKT blot in IOSE 166a panel in Fig. 3A; T-AKT blots in OSEC2 and IOSE166a panels in Fig. 3B.
Journal of Cell Science contacted Dr Auersperg, the corresponding author, to request the original blots, but unfortunately the blots were no longer available. As is our standard practice in such cases, without the original full blots to support the results shown in these figures, the journal referred the matter to the Vice-President, Research and Innovation at The University of British Columbia (UBC), asking them to provide further information and a recommendation as to the next steps.
As part of the case summary provided to UBC, the journal included a visual demonstration of the similarities between the blots in question using the open-source software ‘Forensically’. It is important to note that the software was not used to detect similarities between blots, but rather to highlight them.
UBC contacted the journal upon conclusion of its investigation and stated that no misconduct had occurred. The journal requested the report to understand how the committee arrived at this conclusion, stating that we would have expected the report to comment on the soundness of the conclusions in the paper. UBC provided the report, which showed that the committee had focused its investigation on the use of the Forensically software tool. The report concluded that “…there is no basis for concluding that the blots in question were duplicated or improperly altered.” Their recommendation was: “Care should be taken whenever a report of duplication has resulted from analysis using the current version of the Forensically software.” Furthermore, UBC stated: “It is not the University's place to comment on [the soundness and trustworthiness of the data] – we leave those determinations to the expertise of peer reviewers, editorial boards and the academic community more generally.”
The journal clarified that Forensically was not used to detect any potential duplications, and reiterated its expectations that an institutional investigation would indeed comment on the validity of the data, highlighting relevant sections from the COPE (Committee on Publication Ethics) website; Journal of Cell Science and its publisher, The Company of Biologists, are members of COPE. The journal requested that UBC conduct a more-detailed investigation into the issues raised about this paper, but UBC declined.
The journal therefore referred the matter to an external expert, requesting an opinion on the soundness of the conclusions of the paper by Astanahe et al., considering that the blots in question are unreliable. Our expert advisor expressed the opinion that the conclusions are possibly still sound, noting two additional papers published subsequently that report results similar to those of the paper in question.
The journal also contacted all of the co-authors of the paper, and although there was some initial correspondence, further attempts to communicate with them went unanswered.
The journal is therefore publishing this Expression of Concern to make readers aware of these issues and our efforts to resolve them.