by Fillip; posted on 17 July 2014
We started the CRISPRflydesign website to openly share results of our research with everyone. As the project developed we also wrote up a manuscript, which was first posted on the pre-print server biorxiv.org and was then published in PNAS in a modified form (open access, link). Self-publishing on the web and posting manuscripts on pre-print servers is a great way to foster Open Science, which aims to make the results of scientific research available to all levels of society.
Below I outline some of the reasons why I believe that sharing information on the web is an ideal way to complement traditional publishing of peer-reviewed papers, as it addresses many of the shortcomings of the traditional publishing model.
Instant sharing of results
One great aspect of self-publishing is that we can decide when information becomes publicly available. Once we think a post is ready for sharing we press ‘publish’ and a few seconds later everyone with an internet connection can view it. It is a liberating feeling. Compare that with traditional publishing methods. Once the manuscript is written you submit it to your journal of choice. Then the long wait starts. It might take several days or weeks for the journal to decide whether it regards you work worthy of peer review. If it makes that hurdle, review is likely to take one or two months. And more often than not reviewers will ask for additional experiments and analysis that will further delay publication. You can count yourself lucky if your paper is published 6 months after you began the writing process.
The long time it takes until the paper is published comes with some benefits. During that time the paper has been read by several experts, has gone through multiple rounds of editing and is eventually professionally formatted and typeset. But it also comes at a huge cost. For many months nobody knows about the results of your work. Nobody can build on them, test your new hypotheses or use the material you have generated. You also won’t get any credit for your work. That is not good for science and not good for scientists.
I hope that in the future researchers will routinely use different formats to communicate their work. Self-publishing via personal websites, services such as Figshare or F1000 Posters and pre-print servers is ideal to rapidly communicate – often preliminary – results. And publication of peer-reviewed papers is great to describe a whole body of work and present it in the context of the respective field.
Most academic research is funded by public money and I believe that the public therefore has the right to access the results of that work. Unfortunately, much of the academic literature remains hidden behind paywalls. This not only disadvantages the people who ultimately pay for it, but also restricts sharing information between scientists themselves (and isn’t that what publishing is there for in the first place?).
The call for open access has gained traction over the past few years, but much remains to be done. It is important to realise that it is not the publishing industry that is to blame for the current situation. Most publishers are here to make money, and they do that spectacularly well (profit margins of publishers such as Elsevier and Springer exceed 30%). It is us scientists that make the system work as it does. It is us who decide where to publish our work and whether to make it publicly available or place it behind paywalls. Most journals, whether subscription based or not, allow posting of manuscripts on pre-print servers and doing so is one way to make sure the information is freely available (list of journals policies here).
Data presentation in flexible formats
Towards more meaningful peer review
Peer review is in principle a great way to evaluate science. However, pre-publication single-blind peer review by three to four reviewers, as is the current standard, does a suboptimal job. The key problems are that evaluation is done by a small number of people and before the results can be tested by other scientists. The low N number of referees makes evaluation vulnerable to oversight and personal bias. This is enhanced by the fact that reviewers are usually chosen from a pool of close colleagues, which often have conflicting interests or are friends or foes of the authors. Furthermore, how much impact a study will have in the future remains guess-work. Nevertheless impact is one of the prime criteria for acceptance at (glam-)journals.
To evaluate research more fairly and extensively would require that peer-review is performed by more people. In principle, pre-prints would be a great way to do so, as the manuscript is in the public domain and everyone can freely comment on it. Unfortunately, this is not working yet. Despite being viewed more than three thousand times our pre-print did not attract a single comment. The same is true for most other pre-prints on Biorxiv. We did receive some very helpful emails, but people seem reluctant to publicly post comments. One likely reason is that people are very busy and view commenting as a waste of time, as there is no credit to be gained from it. If we want to change the culture of how we discuss research online it will be important that there is some form of reward to be gained from providing valuable insights and feedback. Perhaps grant agencies should routinely ask for a description of how the applicant has contributed to the scientific discussion online? In the meantime, posting content online is merely the first step towards a better peer-review system.
Sharing unpublished results and material has been great fun so far and has led to new collaborations, many interesting conversations and has overall increased the impact of our work. But of course it is not without risk. Molecular biology is a competitive field and putting out unpublished observations brings the risk of being scooped. However, I believe the risk is minimal if you think carefully about what information you want to share at what time and is easily outweigh by the benefits. We certainly will continue our Open Science approach for crisprflydesign.