Nature is retracting a 2010 paper by a team from Princeton and Drexel on the workings of Plasmodium falciparum, which causes malaria in people. How that came about seems to have been a winding road.
The article — a research letter — titled “Branched tricarboxylic acid metabolism in Plasmodium falciparum,” came from the Princeton lab of Manuel Llinás. It purported to find that:
… tricarboxylic acid metabolism in the human malaria parasite Plasmodium falciparum is largely disconnected from glycolysis and is organized along a fundamentally different architecture from the canonical textbook pathway. We find that this pathway is not cyclic, but rather is a branched structure in which the major carbon sources are the amino acids glutamate and glutamine. As a consequence of this branched architecture, several reactions must run in the reverse of the standard direction, thereby generating two-carbon units in the form of acetyl-coenzyme A. We further show that glutamine-derived acetyl-coenzyme A is used for histone acetylation, whereas glucose-derived acetyl-coenzyme A is used to acetylate amino sugars. Thus, the parasite has evolved two independent production mechanisms for acetyl-coenzyme A with different biological functions. These results significantly clarify our understanding of the Plasmodium metabolic network and highlight the ability of altered variants of central carbon metabolism to arise in response to unique environments.
That clarity soon clouded over. In January 2011, the group issued the following correction:
The samples used for histone proteomics described in this Letter were inadvertently switched, such that the U-13C-glucose and U-13C-15N-glutamine data were inverted. The plots in Fig. 2b and the spectra in Supplementary Fig. 3 have been modified to reflect this. The corrected results demonstrate that 13C-labelling of histone acetyl groups occurs only in cells grown on 13C-glucose and not on 13C-glutamine. Therefore, glucose is the primary source of the acetyl units used for both amino sugar biosynthesis and nuclear protein acetylation. Although U-13C-15N-glutamine does give rise to labelled acetyl-CoA, its localization and function remain unclear. The model presented in Fig. 4 has been modified to reflect these facts, which do not alter the paper’s main conclusions about TCA cycle architecture. The corrected Figs 2b and 4 are shown below. The authors apologize for this error.
Now comes the retraction, which calls the whole thing off, although not before the article was cited 65 times, according to Thomson Scientific’s Web of Knowledge:
We retract this Letter, which reported both reductive and oxidative tricarboxylic acid (TCA) metabolism in Plasmodium falciparum parasites (‘branched TCA metabolism’). The data for metabolic labelling of TCA intermediates remain reliable, but we have come to realize that the interpretation presented is incorrect. Although there is both reductive and oxidative TCA cycle flux in P. falciparum-infected red blood cell (RBC) cultures (as we reported), new data from the Llinás and Vaidya groups (manuscript in preparation) suggests that the reductive flux occurs primarily in the RBCs and not in the parasite itself. Specifically, we have used new enrichment strategies for the parasitized RBCs that enhance our ability to measure P. falciparum-infected RBC metabolic activity without excessive interference from surrounding uninfected RBCs. On feeding 13C5 glutamine, we measured both 13C2-succinate and 13C4-succinate in the infected RBCs, demonstrating that TCA metabolism in blood-stage P. falciparum is not branched but primarily oxidative (cyclic).
As it happens, there’s more news from Nature. The journal/publisher has announced new steps to “improve transparency and reproducibility” — the crux of which appears to be a checklist. According to a press release:
New measures to improve the consistency and quality of reporting in the life sciences research published in Nature and the Nature Research journals are being introduced. An Editorial [which RW readers ought to take a look at] in Nature highlights the problems that result from publishers exerting insufficient scrutiny and failing to publish enough information for other researchers to assess the reliability of results. It also announces measures to improve procedures at the Nature journals, including systematically ensuring the reporting of key methodological details, increasing the space given to methods sections and examining statistics more closely.
Central to the initiative is a checklist that will prompt authors to disclose technical and statistical information in their submissions and encourage referees to consider aspects important for research reproducibility. It focuses on a small number of often-incompletely reported elements of experimental and analytical design that are crucial to the interpretation of research results; it also consolidates several existing policies about data deposition and presentation. In addition, the Nature journals will employ statisticians as consultants on certain papers, abolish length restrictions on methods sections and encourage authors to provide data underlying figures in papers.
The editorial concludes that “tackling these issues is a long-term endeavour that will require the commitment of funders, institutions, researchers and publishers” and urges others to “do whatever they can to improve research reproducibility”.
Reproducibility, of course, is an important and growing concern in science, with projects such as the Reproducibility Initiative and the Open Science Framework launching recently. We’ll of course keep an eye on Nature‘s efforts.
Hallelujah. After decades of forcing authors to cut their methods sections to minuscule lengths, the journals are finally realizing that the manner in which experiments are performed is important. This has been a real problem in the most prominent journals. In many cases I have had to call or email the authors just to get the critical details. Fortunately, most are very cooperative. I did, however, end up having to publish results that failed to replicate an earlier study, but could not give much of an interpretation because the previous authors provided neither the genetic background of their subjects nor the solutions they used for experiments (thanks, Cell Press).
Complete agree. Appropriate methodological detail is a fundamental issue of the very philosophical basis of science – that findings should be reproducible if they are correct, and you can’t have reproduction without any details of what actually happened! Of course, the science world is still catching up on this of reproduction too….
Reblogged this on The Firewall and commented:
After retraction, Nature Magazine pledges to “improve consistency and quality of reporting”
But on the other hand, I have another (current) “example”:http://raphazlab.wordpress.com/2013/04/23/nature-publishing-group-new-editorial-policies-reality-check/ of a complete failure of Nature Materials /Nature chief Editor to actually enforce its existing data sharing policy
Update regarding my comment above: Nature Materials has had a welcome change of mind and he is now saying that the data will be shared within two weeks. Let’s see…
Why allow the correction in the first place? There was a mistake, which necessitated changes of a scale sufficient to warrant fresh review. Instead, we have the usual “defensive correction”.
Journals need to bite the bullet here. A retraction does not damage the brand, instead it upholds it, because it demonstrates the integrity of the journal. The new Nature guidelines are a reaction to a very serious problem. However, as the experience of Raphaël Lévy and stpnrazr shows, policy and practice clearly have a long way to go before they are even in sight of each other. Glorious new rules are of no use if the existing ones are not adhered to; what confidence can there be that the new rules will be enforced?
I am reminded of the emperor who forgot to get dressed.
Nature have a strong track record in mishandling concerns over research validity, despite robustly worded editorials. In their editorial from 2010 (Under suspicion) they stated:
‘…when we receive a complaint, we first do our own tests on the figures to see whether the charges have merit. We also take a careful look at the paper as a whole….. if most of the figures are problematic, we will strongly urge the authors to retract the paper…’
A year later we had:
DNA demethylation in hormone-induced transcriptional derepression.
Kim MS, Kondo T, Takada I, Youn MY, Yamamoto Y, Takahashi S, Matsumoto T, Fujiyama S, Shirode Y, Yamaoka I, Kitagawa H, Takeyama K, Shibuya H, Ohtake F, Kato S.
Erratum in Nature. 2011 Dec 1;480(7375):132.
In fact, most of the figures were clearly falsified – a fact that could be seen by anyone examining the published paper for half an hour – but it was only through the persistence of external commentators (11jigen and possibly others) that the paper was finally recognised to be completely invalid (Retraction in Nature. 2012 Jun 14;486(7402):280.)
That 2010 editorial also says that Nature strongly urges retraction ‘… even if the paper’s main conclusions have been verified independently by other labs. The logic is that the published paper did not accurately reflect the data as they were collected.’
Yet you have the example reported by Raphael Levy, where independent verification (rather than supplying of the original data) is clearly supported by the Nat Materials Editor to be an acceptable way of dealing with serious concerns.
The evidence indicates Nature have become practised masters at ignoring or correcting invalid papers, and that they consistently fail to put their own policies into practice. This is lamentable, because Nature should be setting the standard for other journals, and as importantly academic institutions. Any university dealing with concerns over misconduct can see that Nature are very undemanding – look at the McGill / Saleh case (http://www.retractionwatch.com/2013/01/25/mcgill-committee-says-nature-figures-were-intentionally-contrived-and-falsified/) where the university has reported plans to correct a Nature paper containing intentionally falsified images (and has already succeeded to do so with a PNAS paper). Incidentally months have passed and that Nature paper remains as it was; the investigation was done, figures were contrived, so what’s the delay? Is there actually a debate going on? Is someone from Nature belatedly putting their foot down?
SW Lee is another serial corrector (http://www.retractionwatch.com/2012/11/20/image-correction-in-current-biology-for-harvards-sam-lee/). Rather than retracting the paper, Nature published a correction that stated (in part) “We have also been unable to verify without doubt that the image in Supplementary Fig. 9b shows four different mice [rather than four pictures of one mouse from different angles]”. This correction showed mice with tumors that were ulcerating, and some were more than three times the size permitted under animal welfare regulations. None of which is a concern for Nature.
Ahh, but think of the advert: subscribe today, £1 per impact factor point. Nature play the game and make money. Concern about science is discussed in loft editorials, but the aim is to make money, nothing more, nothing less. Worse, we buy into it. It is, after all, our wallets (our libraries are part of our wallet) that provide the cash. In my view one interesting facet of the Stapel case is that it reflects the fact that science, as a community, is equally culpable, something I have recently blogged about.
Few wish to stand up and be counted, because they are scared – scared that it will cost them grants, tenure, promotion. This is fair enough, because the members of panels that deal with grants, tenure and promotion are generally complicit.