I was very relieved that the Open Research London (ORL) meeting on 19 October 2016 went well. Jon Tennant, Ross Mounce and Liz Ing-Simmons established ORL in Jan 2015 but it faded away after a couple of meetings. I’d been thinking for some time that I should start up the group again, but was a bit wary of the work involved. The group’s founders were happy for me to have a go at re-establishing it and I found a willing co-organiser here at the Crick in Martin Jones.
About 55 people turned up to the spectacular new Crick building to hear talks on two publishing innovations and a talk / interactive workshop on open science workflows.
Preprints – bioRxiv
John Inglis is the co-founder of bioRxiv, the preprint service for the life sciences based at Cold Spring Harbor Laboratory (CSHL). He is also the founding Executive Director and Publisher of Cold Spring Harbor Laboratory Press in New York, a not-for-profit publisher of journals, books, and online media in molecular and cellular biology. He spoke about bioRxiv at three years old.
First a definition. John defined a preprint as “a complete but unpublished manuscript yet to be certified by peer review”. The definition is all about what it’s not (i.e. peer-reviewed) but the key to preprints is all about the speed of dissemination of research results.
Preprints are currently distributed under two models. In the first, a manuscript is submitted to a journal which makes it available as a reprint and then puts it into a peer review process. The journal publishes those preprints that pass peer review. The business model is that of the journal – i.e. it is funded by APCs (article processing charges). An example of this model is F1000Research. In the second model manuscripts are submitted to a service dedicated to preprints. There is no fee and no peer review as part of the service. The service is supported by institutions and foundations. An example of this model is arXiv, supported by Cornell University and others. It gets 100,000 submissions p.a.
John explained that CSHL is committed to science communication so bioRxiv was a natural extension to its activities.
Modelled on the arXiv, it is a dedicated, publisher-neutral service. The time seemed right to launch it in 2013 as there was a new enthusiasm for openness, a greater acceptance of digital resources and practices, and increased posting to the quantitative biology section on the arXiv showed that some biologists at least were ready for preprints.
In building bioRxiv the founders aimed for a simple, reliable submission system, and wanted to ensure that authors would have total control over their content (using a variety of licenses).
John Inglis is always careful to talk about bioRxiv in terms of ‘posting’ and ‘distribution’ rather than publishing. There is no peer review component of the service.
Manuscripts submitted to biorXiv are allocated to one of 26 subject categories then checked for scope and format, and subjected to a plagiarism check. Some basic checks are made to confirm that the submission is scientific in nature, is a report of research (rather than just a hypothesis) and whether there are human health implications. In the latter case some further screening is needed.
If all checks are passed then a few hours later it will be posted on the website. The author can post revisions at any time. biorXiv gets information out there very quickly.
In its first three years bioRxiv has posted 6200 manuscripts. Disappointingly not many of them are confirmatory or contradictory results. 25% of them have been revised at least once. There are more submissions in biology than in health.
Submissions come from 2900 institutions, across 82 countries. Submissions are increasing, currently standing at 450 per month.
Authors seem to like it. Transmission of research results is very rapid. They get useful feedback and their preprints are read. 11% of the preprints receive comments and there have been many blogposts and tweets about bioRxiv preprints – 93k tweets in the past 12 months.
Institutions and funders are starting to accept preprints as evidence of productivity. Both NIH and Wellcome are considering their policies on accepting preprints in grant applications. Rockefeller University has accepted preprints as part of a CV. EMBO accepts preprints on CVs for fellowship applications. Many journals will consider preprints for publication as papers, including Nature, Science and Cell. PLOS Genetics has appointed editors specifically to search bioRxiv for potential manuscripts to publish in the journal. CSHL has undertaken a good deal of advocacy with learned societies and publishers. Several publishers now offer a “submit your manuscript to bioRxiv” button as part of their submission process.
It’s time to start sending your manuscripts to bioRxiv!
Wellcome Open Research
Robert Kiley is Head of Digital Services at the Wellcome Library and is currently acting as Wellcome’s Open Research Development Lead, responsible for developing a new open research strategy for the Wellcome Trust. Over the past decade Robert has played a leading role in the implementation of the Trust’s open access policy and overseeing the development of the Europe PubMed Central repository. He is also the Trust’s point of contact for eLife. He spoke about the Trust’s latest publishing initiative: Wellcome Open Research.
Just as with bioRxiv, the purpose of WOR is to make research communication faster, and more transparent.
WOR is fast, inclusive (you can publish everything – not just standard research narratives), open, reproducible (data is published alongside the article), transparent (it uses open peer review), and easy (the costs are met directly by Wellcome). The only drawback is that it is only available for outputs from Wellcome-funded researchers.
In early submissions, they have seen a range of publication types and a range of researchers from senior to junior. Submissions have come from a range of institutions.
The open peer review process allows for one of three decisions from each reviewer: approved; approved with reservations; not approved. If a preprint gets two ‘approved’ decisions then it is indexed in PubMed and deposited in ePMC. The is the same process used by F1000Research, which provides publishing services for WOR.
Wellcome hope to attract a range of outputs from a range of researchers. They hope that other funders will in due course emulate their initiative.
Open science workflows
Bianca Kramer and Jeroen Bosman are subject specialist librarians from Utrecht University who have considerable expertise in scholarly communication and research workflow tools. Together they led the global survey in Innovations in Scholarly Communication. Their talk, Open Science workflows: Tools and practices to make your research more open, reproducible and efficient promised us a glimpse of the future.
They started by showing some diagrams of research workflows – nice well-behaved cycles with boxes and arrows. But life is not so simple. To make the workflow more realistic Jeroen added cycles within cycles, put in some dead ends and some repeats. By the end it looked more like a game of snakes and ladders.
We then learnt a little about the survey of tools that Bianca and Jeroen had made, and how they had categorised different tools as either:
- Traditional (same as print era)
- Modern (internet era)
- Innovative (social media, collaborative aspect)
- Experimental (new tools, developing tools, startups)
Lest we become too focused on technology and tools though they emphasised that research workflows are less about tools and more about work practices and people.
Bianca and Jeroen are nothing if not practical and they are happiest when considering reality instead of abstractions. The remainder of their presentation was in the form of a workshop, an endeavour to define a set of work practices that would constitute an open science workflow, and tools that would support it.
The scale of their ambition for the workshop, and the amount of preparation they had done, astounded me. There were three or stages to this exercise.
Each seat had a small paper button with the name of a research tool on it. We were first invited to review the tool on our own seat, consider its usage and application in open research, and then to discuss this with our neighbour. Around the periphery of the room a number of research practices had been arranged, grouped according to the different stages of the research cycle. We were next asked to affix our paper button to an appropriate research practice. Finally, we were each asked to move one of the research practices that we considered to be part of an open science workflow and place it in a new area at the back of the room, together with any tool we thought would help with the practice. This back wall filled up with open research practices and tools.
As a participant it was quite challenging to grasp all of this, and a bit mind-boggling to assimilate everything that was on the back wall. I suspect that Bianca and Jeroen, who live and breathe this stuff, had a clearer picture than anyone else n the room. It certainly stimulated us to focus on the role that tools can play in open research practices, and on what an open research practice is. I liked the way too that the whole process of the workshop was open and transparent. Everyone in the room had a sense of the task we’d been set and of how the solution to it was emerging – we could see it on the walls. This really was a great model for open research.
Bianca and Jeroen’s mixture of openness and outreach is a great combination.
You can read their own account of the workshop on their blog.
And that was Open Research London, Oct 2016. We’re planning the next one already – expect it to be Jan or Feb 2017. Hope to see you there! Find out more about Open Research London on the website or follow it on Twitter.