I have read a number of really great articles recently, regarding the future of academic publishing, and I have long been excited by the potential of altmetrics as a mechanism to judge the true impact of a research paper. But having spent the last week chatting with my partner (who just happens to be a rising star at one of the largest publishing houses in NYC) about the future of publishing in general, I am convinced that we need to completely reinvent the way we approach presenting and sharing data. The current system of peer-review and publishing in academic journals was perfectly suited to the environment 20 years ago, it was appropriate for the volume of publications and the size of the data sets presented. This is no longer the case.
The sheer size of some data-sets (think next-gen sequencing, or next-gen imaging), and type (3D modelling, video) cannot be presented in print editions of journals. The current solution is to present these files as supplemental data which are not available in print editions, and are (generally) separate to the main document online. Similarly, there are problems related to journal paywalls (for publicly funded research), lack of detailed methods sections (when did we arrive at the point of a single sentence being acceptable for particular methods?), impact factors (see some great blog posts from Stephen Curry and Michael Eisen), author lists and author credit, time between submission and eventual publication, the single-blind peer-review system, and many more. The open-access revolution has begun to address many of these problems (I am a massive fan of the article-level metrics and the comments section – even if this is currently underused, in the PLOS journals), but these journals are still limited to the standard article format (abstract, introduction, methods, results, discussion).
At the risk of oversimplifying and generalizing, labs tend to have a particular research focus (a disease, a protein/transcription factor, a model organism/cell etc), and that lab will build on that research focus over an extended period of time. A lab may publish several smaller papers (or one large paper) that build around a central issue, and then branch out in a new direction. The several smaller papers may be published in different journals with different first authors. The single large paper may take years from inception to publication. As each lab builds upon each publication, wouldn’t it be great to be able to see a story develop from inception to conclusion (and not the “future studies will investigate….” conclusion, but rather “this has led us to a completely new area….” or “we are sufficiently satisfied with our conclusions and have shifted our attention to…”). Wouldn’t it be great to be able to access the entire story from a single destination, and to be able to comment on the work prior to publication? Wouldn’t it be great to see some of the raw data?
What I’m leading to with this discussion is a proposal to change the way science is presented, shared and judged. What if we had a Facebook of science? What if we took something like ResearchGate a step further and uploaded results directly on a pre-publication weekly/monthly/quarterly basis? Each researcher could have their own page which would include a short summary of their research interest, an introduction that puts their research in context, a detailed methods section (this would include the type of protocols used in the lab, step-by-step protocols to allow for reproducibility), and then (well organised!) figures and tables that are regularly updated with both raw data files and condensed figures. Finally, a discussion of the results obtained and what the next step/s will be. Importantly, all of these sections would be open for comments (from registered users). Researchers would have the option of ‘following’ pages/projects that were of interest to them and could receive notifications of updates. The relative impact of the science being conducted could be judged via page views, links to pages, number of followers, this could be judged on both an immediate level (a researcher uploads an incredibly exciting result once), and across the course of their career (a researcher who consistently uploads interesting results).
A system such as this would increase peer-review from 2-3 reviewers, to however many people find the topic interesting. It would allow for the generation of a complex story over a number of years, will still providing clear evidence of progress (especially important for early- and mid-career researchers). It should encourage the publication of both positive and negative results (imagine how many experiments have been repeated by different labs but never published? Think of how much money is wasted because we don’t publish negative results). It should reduce the likelihood of getting ‘scooped’ after spending several years on a single project (or at least reducing the pain to getting scooped on months rather than years work). It should encourage more open collaboration between research groups, and it should encourage uploading of raw data, thus reducing the likelihood of data manipulation.
I think some of the open-access journals have started this revolution, but I also believe that the revolution is only just beginning.