I was just emailing with Vivian Siegel the other day and she was saying that the Public Library of Science needed to get the word out about itself and get people submitting high quality papers to them. I think it's a very important step in changing scientific publication practices - copyright ownership in particular. Check it out at www.plos.prg (I'm probably not inserting the html right but ... it appears I actually did - so simple!) <a href="http://www.plos.org/"><img src="http://www.plos.org/images/FindBanner3.gif" alt="Lead the Next Scientific Revolution. Publish your best work in PLoS Biology. click here" width="433" height="89" border="0"></a> If anyone has any thoughts on the future of scientific publication feel free to post them. Here was what I had to say (a bit long winded but): Harold Varmus gave a talk to some of the students at Rockefeller University which I attended on April 2nd. At this talk he discussed PLoS, which I was already aware of through signing the "Open Letter". I made a simple suggestion regarding the handling of data and he recommended emailing in my suggestion. While I've decided to elaborate, the basic idea is the annotation of data and claims of a paper in a database/computationally useful format as part of paper submission. In your venture to make changes to copyright and access in scientific publication, largely fueled by the online format, it makes sense to consider other features of this medium and how best to take advantage of them. While I'm sure you have considered such issues, different perspectives yield different ideas. One can speculate with reckless abandon about what seems wonderful, but the basis of your venture is a very good one and it should no be risked by overcomplicating the process or adding unnecessary goals. By the same token as a "public library of science" it seems that major goals should include compiling and integrating scientific information. Even if certain things aren't addressable initially, I think it is important to consider part of your mission to facilitate the development and integration of such approaches into the publication process. Recently in the bioinformatics field there has been increasing interest and efforts in extracting information from text in publications. This is actually quite a difficult thing to do and it seems that some simple way of describing typical claims of a paper should be possible (indeed anyone extracting this information must encode it as well). For instance "molecule A interacts with molecule B" or "gene A suppresses/enhances the phenotype of gene B". It is a much better situation for the person writing the paper, who knows the data best, to state the claims as part of the publication process than annotation by others for literature based databases. Generating this kind of a knowledge base is certainly something quite different from what journals typically do and might be best handled not by the journal per se, but some linked entity. As NCBI already has a fair amount of experience in such ventures it would be a reasonable partner in such a venture (other databases and organizations have specific expertise that would be useful as well, such as the gene ontology consortium). However, requiring that one does "annotate" their claims in such a fashion is almost the essence of scientific publication. It even follows somewhat smoothly from the standard practice of sequence submission currently in practice. In the long run one would envisage that it would actually be nice to have an explicit way of stating what the basis of such claims is from raw data -> derived data -> claims. This would be very useful for those looking into a new area initially as they could explicitly see the basis for claims and understand the limitations and caveats involved. More elaborate approaches which linked in protocol information could even possibly be utilized to assess the strengths and weaknesses of various techniques. If one utilized "belief polling" by some body of approved scientists (or weighted various peoples votes in some reasonable fashion such as publications and position) it would be possible to actually study the evolution of scientific understanding - a whole new field which would be quite interesting (ie for computer scientists in terms of efficiency and psychologists in terms of how beliefs affect direction of research). This is related to having open commentary, which was one idea that Dr. Varmus gave during his talk. Reviews are an important part of science for the simple fact that nobody can keep up with the primary literature. Generally it would be wonderful to modularize and integrate the process of writing reviews so the net result was something like a dynamic online textbook. Indeed it seems that reviews published in your journals could be adapted to this purpose by outside parties given the details of your current license. However, a more structured approach would be preferable. If basic subjects were included this could also open up the ability of the lay public to appreciate and utilize the information gathered in scientific research. On your site you mention public access as a goal, but this is an empty one if they cannot understand what they have access to. Furthermore, if annotation of background information discussed in the paper was made part of the process, it would provide a high quality way of pipelining the knowledge held by the scientific community into a format utilizeable in databases and computational studies. Science's STKE is an interesting example of something similar to this. Since I could imagine someone reading this email concluding that I think that the text of a paper is secondary, I would clarify that is not the case. Text itself is still the best form for disseminating information to other scientists. However for computational approaches, which are becoming increasingly important, to utilize the same information requires another format. As the same information is of interest in both cases, encoding the information into that format should probably be coupled to the publication process due to the gain in efficiency of such a practice. One other thought that occurs in relation to the idea of public access to all publicly funded research would require more of a social and administrative shift, but bears mention as well. A lot can be learned by negative results and partial information that is not the full basis for a paper. It would bear consideration to have some repository of all data generated under public grants as others may be able to utilize that data even if those who generate it can not. While it is certainly reasonable to have some significant period of sole owernship of ones own data, there is certainly some data that might be useful to others which may never become publicly available and this seems a waste. This point is certainly tangential to the actual aims of PLoS, but fundamentally related to the sentiment behind it's generation. As stated earlier, I would not suggest over complicating your initial format and aims, but I hope some of these ideas are at least a useful perspective in the years to come generating a powerful public information resource for science. I would also be happy to discuss these issues further if anyone is interested.
