SoylentNews

BlackHole writes:

In January, a team of scientists from the RIKEN Institute in Kobe, Japan, and Harvard University published two high-profile papers in the prestigious scientific journal Nature (paper 1, paper 2), in which they reported the discovery of a simple method for reprogramming somatic cells (e.g., skin cells, blood cells, etc.) to a totipotent state (i.e., like an early-stage embryo, capable of forming a new copy of the donor organism (a.k.a., a "clone")). The method, which they termed "STAP" (for Stimulus-Triggered Acquisition of Pluripotency) involves using a mildly acidic solution to stress cells taken from any of a variety of tissues (skin, blood, etc.). After stressing, the cells are grown under standard culture conditions for several days, at which point, with no further intervention, the cells become totipotent.

The discovery was celebrated for its broad clinical potential (the resulting cells have all of the capabilities of embryonic stem cells they can be directed to differentiate in culture into any cell type (neurons, bone cells, cardiac cells, etc.)). However, unlike therapies that use embryonic stem cells, because STAP cells are patient-derived, STAP-cell therapies would not carry a risk of transplant rejection. Also, STAP cells are free of the ethical and logistical (i.e., limited supply) issues that plague embryonic stem cell methods.

The reported technique is amazingly fast, shockingly efficient, astoundingly simple, and... wait for it... completely unreproducible. Both papers have now been formally retracted (retraction 1, retraction 2).

However, the bottom line is that this entire episode has affected research elsewhere:

Scientists in the stem-cell field (like this submitter) will recognize that the news here is not (i) that papers purporting to demonstrate a new method for producing stem cells were rushed through the peer-review process (the Nature editor handling these manuscripts could easily have requested evidence of independent replication before publishing), nor (ii) that a new high-efficiency and supposedly simple cell-reprogramming method is, in fact, irreproducible (high-profile protein-based reprogramming method: yet to be reproduced, high-profile microRNA-based reprogramming method (paywalled, sorry): yet to be reproduced, high-profile (one of Science Magazine's Top-Ten Breakthroughs of 2010) messenger RNA-based reprogramming method : yet to be reproduced (PDF)).

What makes this story unusual is the ferocity of the public's response (in Japanese) to what are, quite frankly, levels of data falsification, data fabrication, and plagiarism that are not atypical in this field (see below). As a direct result of this public outcry, the lead author of both papers, Dr. Haruko Obokata (who, since questions about her work first arose in February, has been hospitalized because, according to her lawyer, "her mental and physical condition is unstable") was formally investigated by a committee established by RIKEN comprising senior scientists. In May, the investigative committee found Dr. Obokata guilty of three counts of scientific misconduct. As a telling, and also borderline-farcical aside, during the investigatory committee's investigation of, among other things, alleged image manipulation in one of the papers, an investigative committee was formed to investigate alleged image manipulation in published papers authored by the chair of the original investigative committee, who, unsurprisingly, was forced to step down from his chairmanship as a result. As perhaps the clearest reflection of the overall state of the field, the committee chair was replaced, not by another scientist, but by a lawyer, as apparently no trustworthy scientists could be found.

To restore public trust, the RIKEN Institute tasked an outside panel of experts with (i) investigating the culture at the Center for Developmental Biology (CDB) at RIKEN, where the STAP-cell work took place, and (ii) recommending any policy or structural changes that could be made to help ensure that the science produced at the CDB meets an acceptable level of integrity moving forward. The panel recently concluded that the best course of action would be for the CDB to be "dissolved as soon as possible".

Furthermore, the STAP-cell fiasco has placed in jeopardy RIKEN's bid to be named a Special National Research and Development Corporation by the Japanese government (as well as the increased funding that comes with that special designation).

The aforementioned public outcry has so far been largely limited to Japan. There has yet been no indication of whether Harvard will initiate its own investigation or take any other action in this matter (Dr. Charles Vacanti, of ear-mouse fame, who is affiliated with Harvard, is an author of both retracted papers).

There are many other interesting and relevant aspects of this story that shed light on how things went so wrong, including (i) many pages of Dr. Obokata's doctoral dissertation appear to have been copied and pasted from the website of the U.S. National Institutes of Health, (ii) requirements normally associated with RIKEN's hiring practices (e.g., interviews conducted in English) were disregarded in the hiring of Dr. Obokata, (iii) RIKEN will allow Dr. Obokata to attempt to replicate her experiments, but she will be monitored by video surveillance.

This episode has cast a spotlight on the shortcomings of publicly funded biomedical research, and raises a number of important questions. Starting from the assumption that the primary goal of biomedical research is to improve peoples' health (as opposed to fundamental biology research, which has as its primary goal the generation of knowledge):

1. In light of the following three points, is it a good idea to devote public resources to biomedical research at nonprofit institutions (e.g., universities and research centers like RIKEN)?
a. While researchers at for-profit companies can face financial incentives to generate positive results, researchers at nonprofit institutions also face financial incentives to generate the positive results required to publish high-profile papers (tenure, revenue from patent-licensing, opportunities to engage in outside commercial activity (e.g., consulting, start-up companies, etc.) are all often directly linked to the publication of high-profile papers).
b. While regulatory authorities (such as the Food and Drug Administration in the U.S.) provide strict oversight of clinical research (which constitutes the vast majority of research conducted at for-profit bio-tech and pharmaceutical companies), there is little or no meaningful independent oversight of non-clinical research (which constitutes the vast majority of research conducted at non-profit institutions).
c. There is generally no requirement attached to public funding of non-clinical biomedical research that the funded research be advanced to the point of actually improving anyone's health (e.g., clinical trials). Instead, a successful endpoint is defined as a high-profile publication in a prestigious journal.

2. Can we decouple public funding of biomedical research from potentially corrupting influences, perhaps by avoiding performance-based metrics that are highly susceptible to gaming/fraud (number of publications, publishing in high-impact journals, etc.) when determining the allocation of funding? If so, what metrics, if any, should we use to determine which scientists get public funding, and how much they get (again, focusing on biomedical research, but recognizing that the same principles likely also apply to other fields of applied/commercializable research)?

3. Can we improve the peer-review system, perhaps by requiring independent replication as a matter of course, and/or by requiring papers that have not been independently replicated to carry a disclaimer to that effect (like the albeit often disregarded requirement of a conflict-of-interest disclosure statement)? Would a scientific journal that unilaterally adopts these practices thrive?

4. To disincentivize scientific misconduct, should we try to help the public understand the importance of adhering to the following basic tenets of scientific publication: a. all authors share responsibility for the entire paper, b. irreproducibility of a published method is grounds for retraction, independent of the integrity of the data (i.e., whether or not there is specific evidence of misconduct; think Fleischmann and Pons' cold fusion), and c. fabricated or falsified data in a publication are grounds for retraction, independent of the purported reproducibility of the published method.

Finally, before someone says: "Papers were published, the scientific community couldn't reproduce the results, and the papers were retracted, all within six months. The peer-review process worked! Hugs all around!", please consider the following:

Last December, this submitter's lab submitted a grant application to the U.S. National Institutes of Health, in which we proposed work on a new cell-reprogramming method. Our application was rejected in March (after the publication of the STAP-cell papers, but before the integrity of the papers had been seriously questioned) because, as one reviewer put it, "The [...] field continues to move fast [...] A case in point is the recent successes in reprogramming cells just by stressing the cells in culture" (referencing the now-discredited STAP-cell papers). The point here is that, in our current system, scarce public research funding is allocated before erroneous and/or fraudulent papers can be identified, investigated, and retracted. (btw, our resubmission application just received a great score and will hopefully be funded :), but we (and the patients who may eventually benefit from our work) still lost six months due to the initial rejection :(