Linda Miller, executive editor of Nature, agrees that scientists continue to seek publication in prestigious journals to enhance their own standing. They also concentrate on reading the best-regarded ones, precisely because their time is precious. “You want to be directed, to use the best journals as a filtering device,” she says. “I have been an editor for more than 20 years and I have handled a lot of papers. Every single one has been improved by peer review.”

Update: Hilary Spencer and the Nature Precedings team provide a one-year perspective at Nature Network.

Certainly, peer-reviewed literature and scientific meetings in the physical world will remain the main modes of distributing scientific information and informal communication. Yet, communication through virtual-world technology might become a useful supplement to the traditional discourse. The particular strengths of this technology include: its potential to share, review and comment on information, both with the public and one's peers; options that allow users to create and develop unique objects, and presentations to educate and inform others and to display data; and, last but not least, the time and cost of bringing people together within and across disciplines can be reduced.
As with any new technology, there are issues that could have an impact on the usefulness of online communication and its acceptance within the scientific community. Scientists who rely on peer-reviewed data for their work might find Web 2.0's lack of proofreading unacceptable to document research findings. However, we should explore the existing and potential applications of virtual communication for unique ways to discuss ideas, answer questions, educate and debate. Our ability to understand what we can accomplish in online worlds depends on our collective experience with the technology. The more scientists and clinicians who work with and comprehend the applications of virtual worlds for their respective research fields, the sooner we will realize how this technology can be best applied. The next step is to invite ourselves into these online realms, experiment with what they have to offer, and see where our exploration and creativity takes us.

(Stephen T. Huang is at Northwestern University's Feinberg School of Medicine in Chicago, Illinois, USA; Maged N. Kamel Boulos is at the University of Plymouth, UK; Robert P. Dellavalle is at the Denver Veteran's Affairs Medical Center and the University of Colorado Denver, Denver, Colorado, USA.)

The Editor's summary of this paper: Even in clonal populations of cells, there is significant phenotypic variation from cell to cell. This could reflect the 'noise' inherent in gene expression: or the various cell states could represent stable phenotypic variants. Chang et al. analysed the behaviour of an 'outlier' in clonal populations of mouse haematoipoietic stem cells that had very high expressions of the stem cell marker Sca-1 and found that outliers possessed distinct transcriptomes. Though the transcriptomes eventually reverted back to that of the median cells, while they differed they could drive the cells to express characteristics of distinct cell fates. Thus clonal heterogeneity of gene expression may not be due to noise in the expression of individual genes, but rather is a manifestation of metastable states of a slowly fluctuating transcriptome. These fluctuations may govern the reversible, stochastic priming of multipotent progenitor cells in cell fate decision.

The Nature Reports Stem Cells article provides the details of what the peer-reviewers thought of this paper when it was submitted, and how the authors responded. The initial view of the three peer-reviewers can be summarized as follows:

Reviewer 1 In summary, we believe that this paper reports a novel and important biological mechanism for differentiation: it gives evidence that slow stochastic variations in stem cell state last for a long time and result in different fates. It relates the recent finding of slow fluctuations in human protein levels to the biological outcome of cell fate, and finds long lasting differences in transcriptomes in different subpopulations. It is an excellent choice for Nature, provided that the comments are addressed.

Reviewer 2 Although the phenomenon described is immensely interesting and the idea of heterogeneity being retained within even clonal populations of cells is plausible, the authors merely describe this phenomenon and in some instances, do not provide conclusive data to support their interpretation. If a mechanism was determined this would definitely aid in its novelty and interest. In its current form, I believe this manuscript should not be accepted at Nature.

Reviewer 3 This is a very timely and interesting paper, that should be of interest to a broad range of researchers. However, I have some serious concerns over the modeling.

The details of the reviewers' concerns, and the authors' response, can be read at Nature Reports Stem Cells,

You can also read a set of question-and-answer sessions between editor Monya Baker and several of the peer-reviewers of the recent Nature Insight on Regenerative Medicine, here on The Niche, the blog of Nature Reports Stem Cells.

At first glance it seems to be a shining example of the scientific method in action. Two papers published by biochemist Homme Hellinga and his students at Duke University Medical Center in Durham, North Carolina, claimed a breakthrough in rational enzyme design. Last year, another chemist found that Hellinga's enzymes didn't actually work, which led to the retraction of the two papers this February (see page 275 of this issue and an earlier story on 13 February). Then, this March, a third group published research showing that rational enzyme design really is possible. All has ended happily, it seems, with the field marching forward in triumph.
But examined more closely, the episode reveals some less than happy aspects of science as it is actually practised. For example, the problems with Hellinga's enzymes were identified by John Richard at the State University of New York in Buffalo, who hoped to use the proteins in his own work. In effect, Richard and his two co-workers wasted seven months and tens of thousands of dollars failing to reproduce the results from Hellinga's lab. Richard's subsequent efforts to correct the scientific record thus came at considerable cost, with no discernable benefit to his own career.
This is a perennial problem in science. Many researchers who come across non-reproducible work save themselves extra hassle and money by simply not pursuing it further. Meanwhile, those who refuse to let it go — like Richard — gain nothing.

--donate to funding agencies. It happens in many countries of the world. As long as the agency is committed to assign that money in a transparent manner, including, possibly, peer review, that should be fine.
--donate to individual institutes. In this way, taxpayers may be exercising a little peer-review power. Less troublesome, perhaps, if the institute acts, internally, as the agency above. Still, it is not obvious why institute A should be so much better than institute B. Maybe the cleverest scientist, with the best idea right now, is in institute B.
--donate to a specific project. Here the taxpayers are exercising bolder peer-review powers, and that raises a red flag

Philip Machanick writes that the "problem with blogs is that there is no way for an outsider to know which are reasonably careful creations of informed scientists, which are opinions of the scientifically illiterate, and which are astroturf creations designed to confuse critics of science that is in conflict with an industry. Given that terrain, I would rather have something like RealClimate than not: it helps to balance things out. Errors tend to be corrected quickly here as a consequence of a large informed readership (even if it is sometimes annoying that you get drive-by ignoramuses who don’t benefit from getting their misconceptions answered). He goes on to suggest that a site such as RealClimate is of value in providing a forum for rebuttals of peer-reviewed science in order to develop a consensus on whether a "formally published rebuttal is worth the effort".

Ray Ladbury's view: "When it comes to peer reviewed papers, one has to presume the reader will have a minimum level of familiarity with the subject matter. One also presumes that the reader will have a day job, and so the question becomes whether the information in the paper is of sufficient interest to the average scientist in the community to say, “Hey, take a look at this. It looks mostly correct to me and has some interesting information/insights/methods…” This is not in any way the gold standard in the sciences. The gold standard comes when the community as a whole says, “Hey, cool, I can use this.” The paper is cited. The techniques are used. Science advances. Eventually, what was in the paper becomes part of the tacit knowledge assumed by reviewers.
The tacit knowledge one can presume for a blog like Realclimate is much lower. One presumes there is an interest in the subject–why else would the reader be perusing the blog. One presumes at least a passing acquaintance with the scientific method and maybe some familiarity with basic results like conservation of energy, etc. One could perhaps assume that the average reader has taken the time to acquaint him- or herself with material to which one is vectored via the “Start Here” button–although this is far from Universal.
For the average newspaper reader of a science story, the tacit knowledge is nearly nonexistent–or worse, wrong. And then we have the blogosphere, where information density is at best, rarified and often toxic.......in an information economy, it seems that all too many readers and journalists are content to remain paupers."

Bernd Pulverer, Chief Editor of Nature Cell Biology, has kindly agreed to my request to write a guest post to respond, and to elaborate on the journal's practices and policies:

As ‘Regular scientist’ points out, we do strive to publish rather well-developed studies in Nature Cell Biology. It is interesting to read the concern about the supplementary information, since we usually hear the opposite complaint: namely that the tight format restrictions applied preclude presentation of complete datasets. In considering submissions, we expect fewer details of molecular mechanistic analysis if the paper reports a highly novel observation. And because the study of cell biology tends to throw up surprising new findings that deserve a wide audience but that cannot be developed in a reasonable timeframe, we decided some years ago to publish a short ‘Brief Communication’ format, which I hope is appreciated by Regular Scientist.
In the Editorial under discussion, we wrote:

Nevertheless, the most important part of the report is assessment of the data: are key experiments or crucial controls missing? Are the data significant and definitive? Are all claims made supported by the data? A black and white model should not be a requirement — in fact it often underestimates biological complexity. Formulate a clear set of recommendations for additions or improvements. Experimental suggestions are important even if an outright rejection is recommended, as it makes for a transparent and constructive report that will allow the author to improve the dataset and select an appropriate target journal. If the overall assessment is positive, do not feel obliged to ask for non-essential experiments. New experiments should add key information, buttress claims or improve existing data, and they must be achievable. Indeed, the perception that referees invariably ask for more data encourages some authors to hold back data in the hope that these will be requested and can then be provided by return post; clearly, this is a futile exercise. Referees and editors alike should take great care that all key issues are raised in the first assessment — draw a line and adhere to it. Subsequent evaluations may raise issues on new data, but raising yet more ways to develop a study is not fair.

We, the Nature Cell Biology editors, agree that one can always ask for more data, and that sometimes data require an unrealistic amount of effort for the information added. However, we entirely disagree that our editorial and review process does not result in the publication of much-improved datasets. The tendency of referees to feel an obligation to ask for some new data, and conversely for authors to hold back data to give referees something obvious to ask for, is an issue that we tried to address in the Editorial.
We certainly do not ask for more data because we, the editors, feel under any compulsion to do so: the simple fact is that a good number of manuscripts submitted to the journal are rather premature, and we cannot justify wasting our referee’s time when we perceive obvious holes in a dataset that the referees would invevitably want to be addressed. We believe that this step is in an author’s interest, since we will not undertake multiple rounds of peer-review without good reason. If requested data are deemed to be unobtainable for technical reasons, or indeed unnecessary, we are always willing to hear an author’s arguments.
Regarding the 1953 Nature paper that Regular Scientist recommends: indeed, the paper is a good read. However, science has become much more complex since then, and the discoveries often more detail-oriented, requiring a more in-depth analysis. At the same time, the tools available have developed tremendously in 55 years. Even in 1953, readers will note that Watson and Crick's conclusions depended on the two following papers in the issue, providing experimental details. The challenge is to set ‘the bar’ correctly to what is achievable and what should be achieved for publication in a top-flight journal. This is what the job of an editor is all about.

"Blogs have created a forum for many people to informally discuss science. They are also a great forum for scientists to provide context to their own work and to the work of others in their field.
And some of the time, that may take the form of criticism. That’s ok. But we do have to be careful in blurring the line too much between peer-reviewed publications and blogs. The first problem is the obvious one. Blog posts are unfiltered, un-reviewed, and often written off the cuff, while journal articles are screened, reviewed, and (should be) meticulously researched. It is far easier to write a criticism of a paper on one’s blog than to write a response and submit it to the journal. The issue isn’t just that no reviewers check the work… the blog author is unlikely to do even close to the amount of research and analysis nor give the wording nearly the same level of consideration as is expected in a paper submitted to a high-quality, peer-reviewed journal.
The other problem is that blog posts are readily accessible to anyone at anytime, both in language, and in the unlicensed nature of the internet. To a huge swath of the public, blog posts are the face of science. Like it or not, bloggers with scientific credentials are like self-appointed ambassadors for science. If we are going to write about our science, we should do it with thought, and we should do it well. That is a standard that myself and many other science bloggers often struggle to meet."

More to follow.