The STAP paper that generated so much excitement in Nature has been subjected to some pretty substantial knocks. Several labs have tried to replicate the experiments from this paper, and no one has consistently succeeded. Also, a detailed protocol was released, but the claims of this protocol contradict those in the published paper. Also, one of the authors on the original STAP paper, has even said that he no longer believes the results of his own paper.
The RIKEN institute, where this research was conducted, convened an internal investigation to determine what went wrong. Even though they do not call for the paper to be retracted, they do conclude that deliberate falsification did occur in the paper. There report can be read, in English, here.
The report examines six problems with the original paper:
1. Unnatural appearance of colored cell parts shown by arrows in d2 and d3 images of Figure 1f.
2. In Figure 1i, lane 3 appears to have been inserted later.
3. A part of the Methods section on karyotyping appears to have been copied from another paper.
4. A part of the procedures described in the Methods section on karyotyping appears to be different from the actual procedures used in the experiment.
5. The images for Figures 2d and 2e appear to be incorrect, and closely resemble images in Dr. Obokata’s PhD dissertation.
The first problem is chalked up to what happens to microscope pictures when they are compressed into JPEG files and sent with an electronic copy of a manuscript. Having had figures sliced, diced, shrunk and compressed, blurred, and converted to black and white after submitting them to journals, I can vouch for Dr. Obokata on this one. Therefore, they do think that this one is a problem.
Problem 2 they think is due to true tampering. Lanes in gels, western, southern and northern blots are sometimes cut and pasted in papers, but Nature, apparently has a policy about this and their policy is that this is a no-no. Also, they conclude that the gel lane pasting “created the illusion that the data of two different gels belonged to only one gel, but may also lead to the danger of misinterpretation of the data.” I think they are completely correct on this one.
Problem 3 was probably a dunderheaded mistake. They think that Dr. Obokata plagiarized the protocol, but in all honesty, it could have simply been the result of being in a hurry and having a deadline that you have to meet to finish your Ph.D. and get your papers submitted by a certain date. To my reading, this one sounds like a lack of sleep and being in a hurry. But honestly lady, couldn’t you have at least cited the other paper from which you took the protocol in the first place?
Problem 4 they think is a simple case of someone else did the work and you didn’t check with them first before including it in the paper. Thus it is an oversight and not a case of falsification. On this one, I think the senior author has to bear a lot of the blame. It’s his butt on the line if the paper has anything wrong in it, and he simply did not read the submitted paper carefully enough before submitting it.
Problem 5 is a pretty flagrant case of bait-and-switch. The original figure in the paper was supposed to be STAP cells made from spleen. However, Dr. Obokata said that these were pictures of bone marrow blood cell-making stem cells instead of spleen stem cells. Also the pictures she substituted came from her Ph.D. dissertation, and were of cells that had not been treated with acid, but had been subjected to shear forces by forcing them through a narrow pipette. This is a different experiment than the one she reported. Also, her statements that she had forgotten that these figures of cells treated completely differently are hard to believe. I think we are justified in calling this one a whopper.
Problem 6 is a mislabeling of two figures of cells that came from the same experiment. It is a classic case of the paper being rewritten before publication, the figures being completely reworked, and the labeling getting all messed up. This one is not falsification but it is negligence.
All in all, the paper is a mess. Whatever might have been observed has been fogged over by fraud, negligence, and too many cooks in the paper-writing kitchen. This sounds like too many people were involved in the preparation of the paper and they did not properly talk to each other. This is a black eye for the Riken Institute, which has done so much very fine work. They are to be commended for speedily convening the investigation and for expeditiously examining the evidence. However, large efforts need to have one clearing house for data and all that data needs to be checked, checked and rechecked after every rewrite and before submission.
I think the papers clearly need to be retracted. The investigation does not make that recommendation, but it is the honorable thing to do under the circumstances.