The article, entitled “Toward a new model of scientific publishing: discussion and a proposal” by Dwight Kravitz and Chris Baker of the National Institute of Mental Health at the US NIH, highlights several serious problems with the scientific publishing machine. In particular, the authors note:

• The average 221 days that it takes to get a paper published delays scientific progress as well as young scientists’ careers
• Considering reviewers’ efforts, time spent by authors on revisions, and publication fees, the cost to publish a single peer-reviewed paper in the authors’ field of neuroscience (which publishes some 2,000 papers annually) is over $4,000
• Designed to help publishers prioritize papers, the peer-review system creates an adversarial relationship between reviewers and authors instead of providing useful feedback to authors about the scientific merit of their work.

Most for-profit publishers argue that the review process chaperoned by their editors ensures high standards of scientific research. Not true, argue Kravitz and Baker. “The system is so stochastic and redundant as to be an active hindrance to the progress of research,” they say. Yet the authors are careful not to blame publishers: “This paper is an indictment of the service that we, as a field, ask them to provide.”

How do the authors propose fixing the system? Start by rethinking a publishing model based on limitations that are ancient history. With no need for physical copies of journals, the cost of paper, printing, and distribution has become irrelevant, so publishers could and should guarantee publication for any paper submitted, they say.

Gary Aston-Jones and David Moorman from the Medical University of South Carolina, Faculty Members in the Cognitive Neuroscience section, rate the paper a ‘Must Read‘, and regard the authors’ proposal as “straightforward and results in an improved experience for the entire scientific community.” Aston-Jones and Moorman summarize it this way:

“Reviewers comment and critique on submissions (reviewed in a double-blind fashion) specifically to improve them, publication is typically guaranteed, and a rating system is imposed based on comments of the reviewers and an editorial board with additional interactive reviews continuing after publication to adjust the rated value of the publication. The authors also include additional details such as ways for categorizing published literature to facilitate interaction with existing and future publications with similar subjects. Reviewers are compensated for their efforts and editors compete to attract the best research. The end result is a rapid, unbiased, and complete means of publication of results with built-in filters to assist readers in navigating the growing number of publications.”

The new publishing process that Kravitz and Baker propose “does not completely demolish the existing system, but streamlines it and optimizes it to leverage the currently available technology.” But, they acknowledge, “the process of reforming the current system of publishing will be long, arduous, and fraught with uncertainty.” Indeed.

Editor’s Note: Stay tuned for our announcement next week about F1000 Research, a repository for original research that tackles many of these issues.

Picture by Evan Bench, http://www.flickr.com/photos/austinevan/1225274637/

So how do scientists keep up with the vast number of articles published in their field? Well, unfortunately, they can’t, and that – if you’ll allow me to go off on a short tangent – is why Faculty of 1000 was created: experts in their respective fields highlight the most important papers and provide opinion on why exactly they’re important.

But I digress – knowing how hard it is to keep up with the literature, how do we know we haven’t missed something important? Could at least a fraction of what doctors think they know to be true then be factually inaccurate? The German-Swiss physician and scientist Philipus A. Paracelsus (1493-1541) once said, “Medicine is not only a science; it is also an art.” With this in mind, what part do assumptions play in medical and scientific knowledge?

Amaral questions this in his evaluation of a paper examining how 28 cancer experts differed in their understanding of the complex biological phenomenon of cancer metastasis, a process that has enormous medical importance. The authors of the study found that, despite agreeing on the individual steps of metastasis, no two expert-proposed scenarios were identical.

This led Amaral to question whether we should be so confident in what we think is factual information:

We believe that we know what others have claimed to have discovered, and that any important developments affecting our research area will eventually become known to us. In an elegant and original manner, Divoli et al. demonstrate that those beliefs are not supported by the evidence.

He goes on to explain that he has come across many an article citing his work in the wrong context, even using it support the opposite claim from the one he made. He says that Divoli’s “work supports the need to revamp the manner in which we publish our research.” We at F1000 agree, and perhaps the answer also lies in the way research is disseminated and shared.

Last year, Arif Jinha from the University of Ottawa estimated that the number of articles published since journals came into existence surpassed 50 million articles in 2009. No wonder Amaral describes biomedical literature as a library of Babel:

…too many potential sources where the nuggets we are looking for could be hidden.

I’ve done a little analysis of the F1000 data: on a per-journal basis I took the number of evaluated articles that have subsequently been retracted in the period 2001 to 2011, divided that by the number of articles evaluated by F1000 in the same period, and multiplied by 1000 (per Fang and Casadevall). Now, the resulting graph of the 2010 F1000 Journal Factor (red) versus F1000 Retraction Index (blue) is a little crowded, but that’s almost an inverse correlation, don’t you think?

F1000 Factor vs Retraction Index

Microb Pathog and Cognition don’t have an F1000 Journal Factor (FFj) before 2009 (which is as far back as we calculate). Cognition however had one retraction out of nine evaluated articles back to 2001, giving it an F1000 Retraction Index of 111. Br J Ophthalmol had one retraction and 14 articles evaluated, making its Index 71.4. The FFjs for FEMS Micro Lett, Mol Endocrinol and Plant Physiol are from 2009, by the way, as they weren’t evaluated in 2010.

Now, there’s obviously a problem with small numbers of retractions, but it’s an interesting graph. And there’s (also obviously) much more one could do with these numbers, more than I have time for right now–which is why I’m going to put the F1000 retraction and article counts in this Google Doc for you to play around with if you like. But, seeing as some of you have asked, here’s another interesting graph. For the journals in the Fang & Casadevall analysis, and those with retractions on F1000, here’s a graph of the FFj (red) against the 2010 Journal Impact Factor. Again, click through for a closer look.

F1000 Journal Factor vs Impact Factor

In 2005, the laboratories of Fred Ausubel and Jorge Vivanco, at Harvard and Colorado State University respectively, published a paper in Nature, Mediation of pathogen resistance by exudation of antimicrobials from roots^{10.1038/nature03356}. Essentially, they showed that metabolites from the roots of our old friend Arabidopsis were released into growth media, protecting the plant from a wide range of bacterial pathogens, while a particular strain of Pseudomonas syringae was (partially) resistant and able to infect the plant, causing disease.

Photo: jbeauchamp on Flickr

Sometimes, it all looks like a house of cards.

You can read more about this particular case at Retraction Watch.

NIH/NLM data

So we might expect to see papers carrying the names of a couple of dozen, if not a couple of hundred, contributors (whether they all should be in the list is another matter) on papers arising from sequencing projects or clinical trials and the like.

But anything biology (or medicine) can do, physics can do better. This morning I was searching for a particular author, with a reasonably common surname and first initial, on PubMed. I stumbled across a paper in Physical Review Letters, with the title of Search for a Heavy Particle Decaying into an Electron and a Muon with the ATLAS Detector in √s=7  TeV pp collisions at the LHC. The paper is full of the language you might expect from such a snappy title, such as

The eµ candidate events are required to have exactly one electron and one muon with opposite charge satisfying the above selection criteria. Furthermore, events have to contain at least one primary vertex reconstructed with more than four associated tracks with p_T > 150 MeV.

And it has, count them, 3044 authors. Three thousand and forty four.

How’s that for stamp collecting?

The lead scientist, Paul Riley, is one of our F1000 Members, and I made a little film about his lab a couple of months ago. He did very well not to let slip about the paper to me, especially considering that the manuscript at the time had been with Nature for an entire year! Yesterday at the press conference I talked to a couple of his lab members that I didn’t get to interview in March. Here’s Sveva Bollini, who did the stem cell characterization and Karina Dubé, a grad student who helped her out:

I also cornered Peter Weissberg, the BHF’s Medical Director, who forcefully makes the point that effective new medical treatments come from a framework, a foundation of basic research:

You can watch Paul talking about the findings and their implications at The Scientist website, but I then asked him the length of time it took to get the paper published. Here, he candidly describes how this paper had a rough time through the review process, but in the end the process significantly improved the paper:

[UPDATE: Read the F1000 evaluation by Nadia Rosenthal here.]

Ural Owl, from OpenCage

Back to the awards–there were six: two for research papers (one in biology, one in medicine); one for a case report; one for Editor of the Year; and one each for Open Access advocacy and Open Data. As you might expect from the world’s premiere post-publication peer review service [Us--Ed.] both research papers were selected and evaluated by Faculty of 1000. Because BioMed Central is the leading Open Access publisher, it’s only right that I provide you with free evaluations of these two papers:

The first metazoa living in permanently anoxic conditions

and

Intensive care diaries reduce new onset post traumatic stress disorder following critical illness: a randomised, controlled trial.

Enjoy!

[Update: there are photos of the event on Facebook.] ]]> http://blog.f1000.com/2011/05/06/winning/feed/ 0 http://blog.f1000.com/2011/04/19/seeing-into-the-pore/ http://blog.f1000.com/2011/04/19/seeing-into-the-pore/#comments Tue, 19 Apr 2011 14:33:35 +0000 Richard P. Grant http://blog.the-scientist.com/?p=4215 Continue reading →]]> Back when the world was young I had a deep and abiding interest in the structure of components of the nuclear pore, especially those involved in getting messenger RNA out of the nucleus. The nuclear pore, you’ll recall, is a massive structure in the nuclear membrane of eukaryotes that keeps the DNA in and all that messy cytoplasmic stuff out–except when it needs to let it cross.

And that’s a very interesting problem. How do you allow the transport of cargo you want to move across the nuclear membrane without letting through undesirables? This was, and still is, an area of intense research–Günter Blobel won the Nobel Prize in 1999 for identifying nuclear localization sequences (and donated his prize money to the restoration of Dresden). And while the pathways, including RanGTP and importin ß and all the rest of them are pretty much mapped out, how the FG-repeats of the pore actually work to do this is still a bit puzzling.

Related, as you might expect there’s been quite a bit of work on the molecular structure of the pore. The problem is, this is very difficult to get a handle on, and there’s a gap between the technically impressive yet still rather fuzzy electron microscopy work and high resolution atomic detail of the sort I used to be involved in.

A paper from the Rockefeller University on a technique called polarized fluorescence microscopy^{10.1038/nsmb.2056} a is going to cause quite a stir then, in a good way. There’s a lot of math in there, but the authors say that this will help us “generate a high-resolution map of the entire NPC [nuclear pore complex],” with the aim of, eventually, measuring “nucleoporin rearrangements during nucleocytoplasmic transport and NPC assembly.” They’ve used the technique to confirm the arrangement of the critical Nup84 complex in pores, and, importantly, they can use the technique in live cells.

Read more about it at the Rockefeller website.

Mathur also has another paper in F1000, on ‘stromule branching’, a response organelles make in response to stress: Plastid stromule branching coincides with contiguous ER dynamics, doi: 10.1104/pp.110.170480. Here he explains a little about stromules and why they’re important:

From Ieda et al., 2010

The groups concerned (say my spies) are working together and sharing reagents and protocols. This does raise the question, however: in the era of internet and cheap data storage, why on earth isn’t it possible to share full methods along with the published paper (Journal of Neuroscience, I’m also looking at you)? One for Peter Murray-Rust to mull over, perhaps.

It’s a little bit more effort on the part of the scientist to begin with (trying to piece together methods from multiple people for a paper was never my favourite task), but in the long run, I can’t see who loses.

Fig 1, Gonon et al.

F1000 Member Noam Ziv evaluates a paper on PLoS ONE that looks at the relationship between misleading conclusions as reported in the media and the misrepresentation or overstatement of the underlying data in the literature itself. It doesn’t make for pretty reading:

we speculate that data misrepresentation in the scientific literature might play a part in the distortion of data into misleading conclusions in the media.

With medical research, this isn’t just an intellectual exercise, of course. The authors look at attention deficit hyperactivity disorder (ADHD), and draw the following conclusion (my emphasis):

Unfortunately, data misrepresentation biases the scientific evidence in favor of the first position stating that ADHD is primarily caused by biological factors. Therefore, this misrepresentation does have social consequences regarding ADHD management: it favors medical interventions over prevention and psycho-social interventions.

This ties in with Sir Iain Chalmers’ attack on the way biomedical research is done. It goes beyond simple non-publication of negative (or ‘disappointing’) results, but into the territory of stretching data to say what they don’t want to say. Having said that, it is perhaps understandable, given the pressures of funding and to publish in as high profile journal as possible, that people will overstate their findings.

Understandable, possibly: but when the data so published have real effects on patients’ lives, then we have to ask whether striving to publish into high impact factor journals is in any way ethical.

I’m not convinced it is.

Gonon, F., Bezard, E., & Boraud, T. (2011). Misrepresentation of Neuroscience Data Might Give Rise to Misleading Conclusions in the Media: The Case of Attention Deficit Hyperactivity Disorder PLoS ONE, 6 (1) DOI: 10.1371/journal.pone.0014618

A recent paper in Cancer Cell¹ explores this concept a little, apparently finding that histidine-rich glycoprotein (HRG), which is secreted by the liver, has anti-tumour effects through the re-polarization of macrophages. This paper was evaluated by Guido Serini and Giulio Gabbiani, who study the cell-extracellular matrix interactions that occur during blood vessel formation.

I say ‘apparently’ because Tobias Vogt has raised an interesting question. He wonders if the effects the authors see is a result of host rejection, as it’s human HRG used, in a mouse host. With less than 70% identity between the two sequences, this does seem like a valid concern.

But I don’t know much about this—immunology is generally stuff that happens to other people in my book, unless you’re talking big angry macrophages or, actually, “Smoldering inflammation” (from Serini & Gabbiani’s evaluation). Any immunologists/cancer biologists out there care to chime in?

Rolny, C., Mazzone, M., Tugues, S., Laoui, D., Johansson, I., Coulon, C., Squadrito, M., Segura, I., Li, X., & Knevels, E. (2011). HRG Inhibits Tumor Growth and Metastasis by Inducing Macrophage Polarization and Vessel Normalization through Downregulation of PlGF Cancer Cell, 19 (1), 31-44 DOI: 10.1016/j.ccr.2010.11.009

LeuT 10.2210/pdb2a65/pdb

LeuT is a leucine-transporting NSS orthologue, and there’s some controversy over whether it binds two, or just one, leucine molecule. This is important because determining the mechanism depends on knowing how many molecules of substrate the transporter binds. Leucine is an amino acid—rather large as such things go, but tiny compared with the transporter—and counting how many are bound to at any one time is reasonably non-trivial. The crystal structure of LeuT only had a single leucine substrate molecule, but the authors said the second one had been replaced by a molecular of octylglucoside, a detergent used to keep membrane proteins happy and stable while crystallographers are mucking around with them (and the source of much amusement in my previous life).

A recent paper in Nature (incidentally, from the same lab as published the structure) seemed to nail the problem (one leucine only, Vasili) and was recommended by Gary Rudnick at Yale. But today we’ve published a dissent by Louis De Felice at Virginia Commonwealth University.

Louis says that the conflicting data “require further explanation and experimentation,” which probably means we’re not going to see the end of this one for a while.

Piscitelli, C., Krishnamurthy, H., & Gouaux, E. (2010). Neurotransmitter/sodium symporter orthologue LeuT has a single high-affinity substrate site Nature, 468 (7327), 1129-1132 DOI: 10.1038/nature09581

We humans think we’re so clever, with our hunting, gathering and intensive farming of potatoes for vodka. But the cell biologist’s favourite model, the slime mould Dictyostelium discoideum, has got us beat. This little critter (which, in case you didn’t know, lives as a single-celled organism until times get hard, and then it turns into a Super Slug of Doom) will pack up its own lunch when it goes walkabout.

They then carry bacteria during spore dispersal and can seed a new food crop, which is a major advantage if edible bacteria are lacking at the new site.

Let me say that again: some of the cells in the collective turn into ‘farmers’, gathering bacteria (their food source) and storing it for later (read our News article here.)

The authors describe for the first time a behaviour exhibited by some strains of the facultative metazoan amoeba Dictyostelium discoideum, allowing them to propose the provocative concept that husbandry was invented by such ‘primitive’ organisms.

—Faculty Member Thierry Soldati

Colour me mindblown.

Talin and actin

But at the meeting of the American Society for Cell Biology in December, we heard rumours of a paper in press refuting this finding. And indeed, just before Christmas came the news (again in Nature Cell Biology) that reducing background fluorescence reveals adhesions in 3D matrices. Fraley et al. have responded, although Faculty Member Ken Yamada remains unconvinced by their arguments in his evaluation.

Ken, who I interviewed early last year for a The Scientist article on a novel and relatively easy way of looking at three dimensional growth in culture, admits there is a problem for us mortals:

Consequently, readers might seem to be left with a dilemma, because these two excellent research teams present contradictory findings using the same type of 3D collagen gel and the same cell lines.

He does, however, come out pretty strongly on the side of Kubow and Horwitz, saying

If one research group sees and quantifies a structure and another group reports that they cannot see it, the structure presumably exists unless it can be shown to be an experimental artifact – which does not appear to be the case here. It will be interesting to learn later what experimental problems encountered by Fraley et al. limited their visibility of 3D adhesions, which may help others perform these types of study in the future.

Reinhard Fässler and Herbert Schiller have also independently recommended the Kubow and Horwitz paper as an F1000 must read.

So what happens now? According to Ken Yamada, the field can now proceed to characterize these structures for their tissue-specificity and functions . And breathe a sigh of relief, presumably.

Gil Smith

I have been a research scientist here in the intramural research program for over 45 years. I came here directly from Brown University where I received a Ph.D. in Biology in 1965. My lifetime research interest is to discover the underlying causes for cancer in the mammalian mammary gland.

Initially, I was drawn to the study of viruses that cause cancer in animals and plants. I began to study the mouse mammary tumor virus (MMTV), which is a RNA beta retrovirus and the causative agent for mammary cancer in MMTV infection-susceptible female mice. MMTV is present in the wild mouse population and causes mammary cancer in these mice. MMTV infection does not kill the infected cell and it is commonly passed to the offspring through the mother’s milk but sometimes, in certain inbred mouse strains, as a proviral gene in germ line DNA.

It became clear to me during my studies that the virus replicated in other cell types in the mouse, but these were not transformed to produce cancers. MMTV must insert a copy of its genome as DNA into the host somatic DNA in order to establish a vegetative life cycle. Over the years it became clear that this event was random and that this activity was essential in the neoplastic transformation of mammary epithelium. It became known as insertional mutation. Since this activity is random in the infected cell genome, it was clear that only certain insertional events could be responsible for neoplastic transformation and in addition expansion of cells containing these inserted proviral copies must represent premalignant and malignant mammary epithelial populations. Thus proviral insertions that were found to activate or otherwise deregulate particular genes commonly in mouse mammary hyperplasia and mammary tumors must identify genes important in transformation of mammary epithelium. These became known as common insertion site (CIS) genes.

In collaboration with Dr Robert Callahan we have identified over 40 genes that fit this category in mouse mammary premalignant and malignant populations. The human homologues of nearly half of these CIS genes are deregulated in human breast cancer. There is a public web site describing these findings. But this is only part of the story. As mentioned above, MMTV transforms mammary epithelium almost exclusively. Therefore there must be some important feature of mammary epithelial biology that promotes this susceptibility. My studies over the last two decades have been focused on this issue.

Mammary epithelial stem cell activity has been demonstrated for over 50 years by tissue transplantation. I proposed roughly 22 years ago that mammary epithelial stem cells existed in mammalian mammary tissue and that they could be characterized by an undifferentiated cellular morphology by electron microscopy. This idea was not initially accepted. Today, however, stem cells are the sine qua non of both regenerative and cancer biology.

My recent studies have sought to demonstrate that a hierarchy of stem cell function exists in the mammalian mammary epithelium and further to show that these activities are dependent upon signals and influences predicated by the tissue microenvironment, which is comprised not only of the stroma and external factors but also the differentiated epithelial cells resident in the organ. To illustrate this, our laboratory has demonstrated that non-mammary cells and cancer cells can contribute normal functioning progeny to regenerating mouse mammary epithelium when mixed with differentiated mammary epithelial cells and placed within the mammary fat pad. Indeed, our recent paper reviewed by the Faculty of 1000 shows that human cancer cells may be reprogrammed in this way as well. The long term goal of these studies is to discern these “reprogramming” signals with the goal of developing a new paradigm for treatment and control of cancer cells in situ.

Now, Quinn is trying to answer the “fundamental unanswered question” in his paper: what exactly was the mechanism that drove the evolutionary loss of these introns?

That simple model is not correct

Figure 1 from M Zahran et al., 2010

rpg: What’s so important about restriction enzymes in general, and what you’ve found specifically?

Petra: Restriction enzymes are crucial for the survival of the cell. They cleave invading foreign DNA at a specific DNA sequence which is recognised among an enormous molar excess of structurally similar non-specific DNA. DNA is not just a stiff double helix, but can be strongly bent.

EcoRV bends DNA before cleaving it and our computer simulation research provides a clear picture of the mechanism by which this bending, clamping and cleavage is achieved. Also, we have been able to show how the intrinsic propensity of specific DNA sequences to bend helps the enzyme be highly selective, and how the protein amplifies this propensity to clamp onto it.

rpg: What’s the wider impact of this research?

Petra: People might not know that DNA is intrinsically bendy, depending on the sequence, and that this is used by enzymes that cleave DNA to bind to it and distort it, hence looking at the recognition process (also) from a DNA perspective. This view can be transferred to DNA processing enzymes in general: the intrinsic DNA dynamics can help those enzymes find their target sites, such as specific sequences or mispaired or damaged DNA .

rpg: How did you get interested?

Petra: The University of Heidelberg has a special research area in catalysis, and research on metalloenzyme function forms part of that. We decided that these restriction enzymes, or DNA scissors, that are so commonly used in molecular biology labs, needed to be understood in detail. We felt that besides investigating the catalytic cleavage mechanism itself, we should learn more about how the enzyme achieves its specificity and elucidate the mechanism of sequence recognition.

The paper, evaluated by B Montgomery Pettitt , is Mechanism of DNA recognition by the restriction enzyme EcoRV and you can read it in full at J Mol Biol.

Just how much is shown by this analysis from researchers at MIT, Cornell University and University College London, with help from the BT Group in a little town you’ve probably never even heard of (Ipswich). The picture represents bidirectional traffic cooked up from 12 billion telephone calls over a one-month period.

Each of the 3042 pixels represents an area 9.5 km by 9.5 km, the opacity dependent on the length of the call time. London sits like a giant medulla, controlling the rest of the country.

For more, read the (Open Access) paper at PLoS, Redrawing the Map of Great Britain from a Network of Human Interactions. Or simply enjoy the graphic.

Science magazine has made the original paper free to view (with the lead author’s personal email address in plain sight. Eek.) Our evaluation is also free to view.

And just for fun, here’s Baruch Blumberg talking about NASA’s Astrobiology Institute, from Web of Stories:

You work ten years on a mission and it blows up before it takes off

Video from Web of Stories; see all of Baruch Blumberg’s stories.

What Felisa Wolfe-Simon and colleagues found was a type of bacterium that can grow (albeit very, very slowly) in surprisingly large amounts of arsenic. That in itself is pretty neat, and it’s really rather a shame that NASA pimped it up to be something else, even calling a press conference to announce a press conference to announce a paper.

But as with most things in science, the initial paper is by no means the full story—and it has already been taken apart quite thoroughly by Rosie Redfield. What’s really interesting, and is likely to have more of an impact on science as a whole, is NASA’s contention that Rosie’s critique can be dismissed because it wasn’t written in a peer-reviewed journal. NASA seems to be really failing to grasp the point here, because scientists (experts in the field, at that) will talk about papers and findings wherever they can, on blogs, on twitter; and only eventually in a peer-reviewed journal. I’m all for opening up communication within the scientific endeavour (as well as between scientists and everyone else), so this kind of discussion is, in my opinion, undoubtedly a good thing.

What [NASA spokesman Dwayne Brown] fails to see or refuses to acknowledge is that Rosie Redfield is a peer, and her blog is peer review. NASA has bungled its presentation of this paper from start to finish. It makes worse by trying to dismiss critiques this way. This is the wrong stuff.

—David Dobbs

I’m quite interested in what NASA would make of F1000 to be honest. A Bacterium That Can Grow by Using Arsenic Instead of Phosphorus has already been evaluated (that’s a free link, by the way) by Laura Landweber at Princeton; it will surely gather more evaluations, and possibly even a dissent or two. This is post-publication peer review, and we’re all part of it.

According to Douglas Erwin, a palaeobiologist at the Smithsonian Institution in Washington DC, such complexity indicates that sponges must have descended from a more advanced ancestor than previously suspected. “This flies in the face of what we think of early metazoan evolution,” says Erwin.

Doug said that when this was published, he was getting calls from colleagues asking if he’d gone mad.

What he was saying is that the idea that modern sponges have descended from a more complex ancestor “flies in the the face of what we think…” Or, as he said to me, “that’s almost certainly wrong.”

Just in case there’s still any doubt: Doug Erwin doesn’t think that sponges evolved from a more complex ancestor. We’re happy to help set the record straight.

Ben Davis at the University of Oxford first evaluated the paper a year ago, a few weeks after its publication. Ben starts off effusively— If this worked it could be marvellous, superb—but then strikes a cautionary note:

this work should be read with some caveats. Try as we might, my group, as well as many colleagues, and I have tried to determine the chemistry described … but questions remain. For example, it is unclear how the authors chemically generated such a large substrate library.

Then of course, just after Science published its Expression of Concern, first Michael Gelb (UoW) and then David Cane (Brown University) published dissenting evaluations. David writes (my emphasis):

My colleagues and I have tried to decipher the chemistry shown in Figure 1 of the main text and in the supplemental material. Many of the indicated reactions seem highly unlikely to occur, and the NMR data showing that some of the structures that were made are confusing and controversial. I would encourage chemical biologists to read this paper and draw your own conclusions.

Which is a pretty strong way of saying “Guys, this doesn’t add up.” Ben replied with tacit agreement.

Dissents are reasonably uncommon at F1000, but for us they are as essential as expressions of concern, and indeed retractions, in the scientific process. (You can find a list of all dissents on F1000 with this handy link: http://bit.ly/F1000dissent.) They drive home the point that science doesn’t not proceed inexorably in quantum increments. Every piece of work; every paper; every table or figure or dataset can and should be challenged. Progress is sometimes backwards, or sideways, or diverted down a B road for a while. Yes, it’s frustrating, and can be utterly disheartening for the scientists involved, but these questions, these dissents and retractions demonstrate that the system works, more or less.

And because we think that this conversation is so important, in the new F1000 website (you knew we had one, right?) we’ve provided the facility for subscribers to make comments on evaluated papers (oh all right then, another handy link for you: http://bit.ly/F1000comment). Any signed-in subscriber can comment on the papers we evaluate, so go on, tell us what you think.

You should probably read the Terms & Conditions first, but they’re not onerous.

Pillar Coral, Florida Keys National Marine Sanctuary

BP effects worse than predicted?
Although many scientists expected this year’s BP oil spill to cause major damage to the Gulf of Mexico, preliminary surveys were cautiously optimistic, causing some to hope deepwater coral “had perhaps dodged an ecological bullet,” according to the New York Times.
No such luck, suggests a submersible robot cruising the seafloor 7 miles southwest of the well, which found widespread devastation. “I have seen many individual dead coral colonies over the years, but I’ve never seen a site full of dead and dying coral colonies,” Charles Fisher, chief scientist on the gulf expedition, told the newspaper. It’s not yet clear whether the oil killed the coral, but Fisher said the circumstantial evidence points to that conclusion. This is only the first nearby coral site he’s surveyed, and he plans to visit others in December.

And the next president of the Donald Danforth Plant Science Center is…
…James C. Carrington, director of the Center for Genome Research and Biocomputing at Oregon State University.

RIP, brucellosis expert
Margaret Meyer, an early pioneer in brucellosis in animals, died at 87 last month from complications due to pulmonary disease, according to the Sacramento Bee. During her four-decade-long career at the University of California, Davis, she traveled the world to investigate and classify the infectious disease, which strikes cattle, bison, and other domestic and wild animals. “She was a leader in publishing information on brucellosis,” Bennie Osburn, dean of the UC Davis School of Veterinary Medicine, told the Bee. “It put her out there as one of the top experts in the world.” Meyer also worked to give women more opportunities in academia, telling stories about how her early professors used to give her twice as much work as male students to try to get her to drop out of science.

Germ-zapper bad for pregnancy?
It’s the start of flu season, causing more people to reach for antibacterial soaps, but new research suggests an ingredient could disrupt the metabolism of estrogen, potentially causing problems in pregnancy. In addition to antibacterial soaps and lotions, the chemical, triclosan, is present in hundreds of other popular products, including socks and toothpaste, but a new Environment International paper shows it can hinder estrogen sulfotransferase, which helps metabolize the hormone and transport it to the developing fetus. “We suspect that makes this substance dangerous in pregnancy if enough of the triclosan gets through to the placenta to affect the enzyme,” author Margaret James, a professor and chairwoman of medicinal chemistry in the University of Florida College of Pharmacy, said in a statement.