By Sally Burn
Two weeks ago we reported on the groundbreaking creation of STAP (stimulus-triggered acquisition of pluripotency) cells by scientists at the RIKEN Center for Developmental Biology in Japan. STAP cells are generated by subjecting specialized cells to physical stress, causing them to reprogram back to a pluripotent state from which they can differentiate into both embryonic and extra-embryonic tissues. The pair of Nature papers [1,2] reporting these findings were received with both excitement and surprise by the scientific community. Excitement because the implications were so fantastic (patient-specific regenerative medicine) and surprise because the system was so amazingly simple. The main stressor used in the experiments was acid; the mature starter cells were subjected to acidic conditions and this was sufficient to send them on the road to reprogramming. At the more positive end of the response spectrum, commentators simply asked: “can it really be that simple?” Responses at the other end were pricklier, with a number of eminent scientists dousing the findings with skepticism.
A number of events have now occurred since publication: a formal investigation by RIKEN into the authenticity of the papers, crowdsourced attempts to repeat the experiments, and successful requests for open access to the papers. The bare bones of the case are this: RIKEN is reacting to concerns raised on blogs by other researchers about the papers plus a 2011 paper by the same author, Dr. Haruko Obokata. These concerns are two-fold: first, that images have been reused in different figures to show different things; secondly, that the results cannot be reproduced by other labs. The first concern is what RIKEN is actually investigating and is potentially explicable (human error). The second issue is not under formal investigation but it is this aspect that has made the news, with reports generally stating that the experiments are flawed and that this is why RIKEN is investigating.
The positive aspects of these events are actually quite exciting. Firstly, freely accessible blog-based discourses between scientists attempting to replicate the studies and sharing their data have sprung up. Secondly, Nature has acquiesced to requests to make the original papers open access. The negative aspect of the fallout, however, leaves a particularly bad taste in my mouth. Perhaps it is just the coverage I’ve seen or the conversations I’ve had, but there is a distinct air of a witch-hunt about the whole affair. The deterioration in opinion from excitement to “I told you so” sanctimony has been rapid. A number of the comments on the STAP reproducibility blogs are pure nit-picking. One of the blogs even holds weekly public opinion polls which serve no scientific purpose on whether we think the data is real or not (64% of respondents felt on the “not” side of the fence at the time of writing).
I will hold my hands up at this point and say that I am biased in hoping that Dr. Obokata will be exonerated. I was enthralled when I read about her: 30 years old and running her own lab at RIKEN?! I read interviews in which she described the pain of working on her research to the exclusion of all else and being met with constant criticism and skepticism. No one seemed to believe her system worked so she repeatedly went back and tried again, adding more controls, building extra layers of evidence. It apparently took five years from the first time Nature saw an earlier incarnation of the paper to them accepting it. But finally a team of peer reviewers accepted the manuscripts and at last, for a short while, her work was finally validated. Overnight Obokata became a global science sensation; STAP cells were everywhere. She became my new role model: an exceptionally successful young independent female scientist. The important part of that accolade is “independent female”. Achieving scientific independence is exceptionally difficult in 2014. The pool of PhD students and postdoctoral researchers in the biological sciences is drastically larger than it was 20-30 years ago. The number of permanent Principal Investigator (PI) positions, however, is not. Moreover, males greatly outnumber females at the PI level (for a great discussion of this see this). There are therefore fewer visible role models for girls getting into STEM. So I am absolutely, unashamedly biased in hoping that Obokata will be exonerated.
If Obokata’s studies are, however, found to be flawed it will result in the loss of a promising role model for women in STEM. It may also damage the public’s perception of the scientific process. The investigation has been reported by many major media outlets. The knee-jerk response by the public may be to view research as unregulated and to conclude that this kind of affair is commonplace – after all, Nature is one of the most esteemed and well known journals. Similar responses by those with involvement in science funding would be an even worse consequence, which is why we need to take measures to reduce the incidences of not just scientific fraud but also the potential for perceived fraud.
Nature is a top-ranked journal and submitted manuscripts undergo strict peer review. Obokata had faced years of skepticism and accordingly had added more and more experiments to strengthen her case. So what could be done differently? One answer lies with the authors themselves as image duplication can occur as a result of human error. It really shouldn’t, particularly not when you’re submitting your career-defining dogma-shaking paper to Nature, but it does – even more so when the stakes are that high. Despite what the public might think, science isn’t performed in a vacuum by moon-suited automatons; it’s usually performed by tired, overworked, underpaid graduate students and postdocs. They are humans and they are likely humans who are running on a diet of black coffee and seminar pizza. They have homes and lives, but they’ve probably not seen those things in the last 16 hours. They mess up. I’ve done it. You, dear reader, have done it. And messing up is even more likely when under pressure (conference poster due in an hour; carrying out novel experiments at midnight to avoid being scooped). What can be done to prevent such mistakes? No single thing will fix the problem. Indeed it may even be the personality of many scientists which increases the risk of such errors: working too hard, too obsessively on a problem to the extent they do not function optimally. However, improving working conditions would certainly help, as would less emphasis on “publish or perish”. The pressure to publish also often underlies actual falsification of data. If you don’t get the data you don’t get the paper and you can kiss goodbye to tenure.
If the image duplications were genuinely due to human error, this brings us to the subject of transparency. Should RIKEN have announced their investigation? By doing so the press has had a field day informing the public that the investigation happened because the science is bogus. Even if Obokata is cleared of any wrongdoing she will always carry the stigma of being accused of fraudulence. Despite this, my answer is: yes. Investigations into scientific malpractice should absolutely be made public. Open access publishing has made great headway in the last decade with bringing science to a wider audience, and we need to do all we can to promote transparency in the scientific process.
In fact, this investigation could end up being a key event in the drive to make science more openly accessible. A very interesting thing happened in the days after the STAP articles were published. Blog forums started in which scientists from around the world shared their experiences of trying to repeat the purportedly simple acid-reprogramming experiments (see particularly PubPeer [3,4] and Paul Knoepfler’s blog). Such openness is rare with unpublished data, but replicating published data opened the door to free communication. At the time of writing the PubPeer page for the first of the two STAP papers had been viewed over 27,000 times. Twitter is also abuzz with commentary on the STAP situation and Reddit joined in when an alleged technician from the lab of Charles Vacanti (the papers’ final author) started discussing the protocol with interested parties.
Thus far no one has managed to fully reproduce Obokata’s results (ten attempts have been reported). Nature has not launched an official inquiry so far but, intriguingly, seems to be conducting its own crowdsourced investigation: “None of ten prominent stem-cell scientists who responded to a questionnaire from Nature‘s news team has had success”. This does not necessarily mean that the original data was fraudulent. Scientific data should, by definition, by reproducible but all too often it’s not the case. We all know of someone who had the most amazing cell system setup but it only works with one batch of media, or the animal experiments affected by external factors (water, food, stress). Again, science is not the sterile logical beast the public perceive it to be. If you perform an experiment twenty times and it works, then you get a new batch of media and it suddenly fails, do you scrap all your data? The answer idealistically is yes but if you are under pressure to publish all too often it will be no. You publish. Or perish…
It may well be that there is something about Obokata’s lab’s setup that isn’t being accurately reproduced elsewhere (the presence of her infamous “good luck” pet turtle perhaps?). For starters, most of the attempted repeats did not use the same starter cells as in the STAP papers. Furthermore, while her experiments were universally reported as “simple”, they are also something she has been expertly honing for many years. Indeed, a co-author, Teruhiko Wakayama, found that he could not create STAP cells outside of Obokata’s lab but he is adamant that he generated them during his time there. In an ideal world external scientists need to go to Obokata’s lab, be trained by her and use her lab resources to reproduce the data.
A final positive outcome of this case is that Professor Paul Knoepfler, curator of the STAP reproducibility blog, took to Twitter to ask Nature to make the STAP papers open access (Nature is usually accessed via an institution subscription or by paying a large fee per article). Nature agreed and now anyone in the world can read the papers for themselves – a victory for open access publishing. A further request by would-be STAP creators is for the authors to release detailed methods; this currently remains unfulfilled. Indeed, Obokata herself has not issued any comment. But then again she is only being investigated for potentially accidental image duplications – a fact being overlooked by many. I sincerely hope that she will be cleared of any wrongdoing, not just because of the otherwise damaging effects on her career and the public perception of science, but also because her findings are just so amazingly groundbreaking. In the words of the X-Files: I want to believe.
Want to stay on top of what’s happening with STAP cells? Create your free feed today.