Tag Archives: Open Research

Open Research Project, first thoughts

Uncategorized, ,

Dr Laurent Gatto is one of the participants in the Office of Scholarly Communication’s Open Research Pilot. He has recently blogged about his first impressions of the pilot. With his permission we have re-blogged it here.

I am proud to be one of the participants in the Wellcome Trust Open Research Project (and here). The call was initially opened in December 2016 and was pitched like this:

Are you in favour of more transparency in research? Are you concerned about research reproducibility? Would you like to get better recognition and credit for all outputs of your research process? Would you like to open up your research and make it more available to others?

If you responded ‘yes’ to any of these questions, we would like to invite you to participate in the Open Research Pilot Project, organised jointly by the Open Research team at the Wellcome Trust and theOffice of Scholarly Communication at the University of Cambridge.

This of course sounded like a great initiative for me and I promptly filed an application.

We had our kick-off meeting on the 27th January, with the aim of getting to know each other and somehow define/clarify some of the objectives of the project. This post summarises my take on it.

Here’s how I introduced myself.

Who are you?

Laurent Gatto, Senior Research Associate in the Department of Biochemistry, physically located in Systems Biology and the Maths Department. SSI fellow and Software/Data Carpentry instructor and generally involved in the Open community in Cambridge, such as OpenConCam and Data Champions initiative.

What is your research about and what kind of data does your research generate?

My area of research is computational biology, with special focus on high-throughput proteomics and integration of different data and annotations. I use raw data produced by third parties, in particular the Cambridge Centre for Proteomics (mass spectrometry data), and produce processed/annotated/interactive data and a lot of software (and also here).

What motivated you to participate in the Pilot?

Improve openness/transparency (and hence reproducibility/rigour) in my research and communication, and participate in improving openness (and hence reproducibility/rigour) more widely.

What kind of outputs are you planning to share? Do you foresee any difficulties in sharing?

My direct outputs are systematically shared openly early on: open source software (before publication), pre-prints, improved data (as data packages). Difficulties, if any, generally stem from collaborators less willing to share early and openly.

A personal take on the project

It is a long project, 2 years, and hence a rather ambitious one, of a unique kind. Hence, we will have to define its overall goals as we go. The continued involvement of the participants over time will play a major role in the project’s success.

What are attainable goals?

It is important to note that there is no funding for the participants. We are driven by a desire to be open, benefit from being open and the visibility that we can gain through the project, and the prospect that the Wellcome Trust will learn from our experience and, implement any lessons learnt. We get to interact with each other and with research support librarians, who will help us throughout the duration of the project. We also commit to sharing of research outputs beyond traditional publications and to engage with the Project, by participating in Project meetings and contributing to Project publications.

A lot of our initial discussions centred around rewards for open research or, actually, lack thereof and perceived associated risks. Indeed, the traditional academic rewarding system and the competitiveness in research leaves little room for reproducibility and openness. It is, I believe, all participants hope that this project will benefit us, in some form or another.

A critical point that is missing is the academic promotion of open research and open researcher, as a way to promote a more rigorous and sound research process and tackle the reproducibility crisis. What should the incentives be? How to make sure that the next generation of academics genuinely value openness and transparency as a foundation of rigorous research?

Some desired outputs

Ideally, I would like that the Wellcome Trust’s famous Research investigator awards to be de facto Open research investigator awards. There’s currently a split (opposition?) between doing research and supporting open science when doing research. In every grant I have written, I had to demonstrate that the team had a track record, or was in a good position to successfully pursue to proposed project. Well, how about demonstrating a track record in being good in opening and sharing science outputs? Every researcher submitting a grant should convincingly demonstrate that they are, have been and/or will be proactive open researcher and openly disseminate all the outputs. By leading by example in the frame of this Open Research Project, this is something that the Wellcome Trust could take away from.

Unfortunately, it is a fact that open science is not on the agenda of many (most?) more senior researchers and that they are neither in a position to be open nor that open science is a priority at all. I find it particularly disheartening that many senior academics (i.e. those that will sit on the panel deciding if I’m worth my next job) consider investing time in open science and the promotion of open science as time wasted of actually doing research. A bit like time for outreach and promotion of science to the wider public is sometimes looked down at, as not being the real stuff.

Another desire is that this project will enable us to influence funders, such as the Wellcome Trust, of course, but also more widely the research councils.

As a concrete example, I would like all grants that are accepted to be openly published beyond the daft layman summary. Published grants after acceptance should include data management plan, the pathway to impact, possibly more, and these could then be used to assess to what extend the project delivered as promised.

This serves at least two purposes. First, it is a way to promote transparency and accountability towards the funder, scientific community and public. Also, it is a great resource for early career researchers. Unless there is specific support in place, writing a first grant is not an easy job, especially given the multitude documents to prepare in addition to the scientific case for support. And even for more experienced researchers, it can’t harm to explore different approaches to grant writing.

Another concrete output is the requirement for a dedicated software management plan for each grant that involves any software development. I certainly consider my software to be equivalent to data and document it as such in my DMPs, but there seems to be a need for clarification.

I believe that I do a pretty decent job in conducting open science: pre-prints, open access, release data, … In the frame of this project, I shall do a better job at promoting open science for its own sake.

I also hope that by bringing some of my projects under the umbrella of the the Open Research Project, I will benefit from a broader dissemination that will, directly or indirectly, be beneficial for my career (see the importance of benefits and rewards above).

Next steps

It is important to make the most out of this unique opportunity. We need to create a momentum, define ambitious goals, and work hard to reach them. But I also think that it is important to get as much input as possible from the community. Nothing beats collective intelligence for such open-ended projects, in particular for open projects.

So please, do not hesitate to comment, discuss on twitter or elsewhere, or email me directly if you have ideas you would like to promote and or discuss.

Published 08 March 2017
Written by Dr Laurent Gatto
Creative Commons License

We are going OPEN – the Open Research experiment has begun!

Uncategorized, ,

There has been much discussion recently about the reproducibility crisis and about the growing distrust among the public in the quality of research. As illustrated in our ‘Case for Open Research’ series of blog posts, one of the main reasons for this is that researchers are currently rewarded for the number of papers they publish in high impact factor journals, and not necessarily for the quality of work that they are doing.

Indeed, Cambridge researchers clearly indicated that the lack of incentives to do anything other than publishing in these types of journals is one of the main blockers discouraging them from adopting a more open research practice.

Joining forces with the Wellcome Trust

The Office of Scholarly Communication started talking about these problem with the Open Research team at the Wellcome Trust. The Wellcome Trust are natural allies, as they have consistently led their researchers towards greater openness. They were one of the first funding bodies to introduce policies on Open Access and on data management and sharing. Now the Wellcome Trust is moving towards proactively supporting Open Research beyond enforcing their compliance requirements.

To promote immediate and transparent research sharing, they have recently launched the Wellcome Open Research platform which allows researchers to submit articles about virtually any research output and get published within a couple of days. The Wellcome Trust is now considering making Open Research one of their strategic priorities.

We quickly realised that we have a lot of shared interests, and joining forces to tackle the problem together made a lot of sense. We came up with the idea to launch the Open Research Pilot Project.

The Open Research Pilot – understanding the barriers to “openness”

We conceived the project as a two year experiment, which would allow us to gain an understanding of what is needed for researchers to share and get credit for all outputs of the research process. These include non-positive results, protocols, source code, presentations and other research outputs beyond the remit of traditional publications.

The Project aims to understand the barriers preventing researchers from sharing (including resource and time implications), as well as what the incentives are. The Project aims to utilise the new Wellcome Open Research publishing platform, together with other channels, to share these outputs.

The invitation to take part in the Pilot was sent to all researchers at Cambridge funded by the Wellcome Trust. Participating researchers had to commit to sharing of research outputs beyond traditional publications and to engage with the Project, by participating in Project meetings and contributing to Project publications.

Is ‘doing the right thing’ enough incentive?

Our biggest question was whether anyone would be willing to participate in the Pilot. We did not offer any incentive other than encouraging researchers to contribute to the greater good. The only support available to those who wanted to take part in the project was that offered by the Wellcome Trust and Cambridge Open Research team members, but there was no financial aid available to prospective participants. We thought that regardless of the outcome, that inviting researchers would be a good exercise to go through – we thought that if no one applied, we would have learnt that doing ‘the right thing’ was not a good enough motivator.

Thankfully, we received several fantastic applications from individual researchers and research groups who demonstrated great interest in and motivation for Open Research. We initially planned to work with two research groups, but given the quality of applications received and passion for Open Research expressed by the applicants, we decided to extend the scope of the project to four research groups. We have selected researchers doing different types of research, with the aim of learning about distinct problems in sharing that are experienced in diverse research disciplines:

  •       Dr Laurent Gatto –is  doing computational biology research, with a special focus on proteomics data. His interest is: How to effectively share research data and the code needed to reproduce them?
  •       Dr David Savage – is researching molecular pathogenesis of the consequences of obesity. His question is: What are the problems with sharing data coming from human participants?
  •       Dr Benjamin Steventon – is a developmental biologist generating and analysing large-scale imaging datasets. He would like to know: Are there image repositories allowing one to share large image datasets in a re-usable way?
  •       Dr Marta Costa and Dr Greg Jefferis (and others) – researchers leading the work on two collaborative projects: Connectomics and Virtual Fly Brain, which will create interactive tools to interrogate Drosophila neural network connections. They would like to understand: What are the issues with sharing complex interactive datasets? How to ensure long-term preservation of complex digital objects?

Motivations

So what motivated these researchers to apply for the project? We asked this question at the application stage and were positively surprised by the altruistic answers that we received. Our researchers were largely driven by a desire to improve the research process. We have seen responses like:

  • “Openness in research, from data and software to publication, is a central pillar of good research.”
  • “I am very concerned (disappointed as a scientist) by the current wave of ‘unreproducible’ and/or ‘irrelevant’ research, and am very passionate about contributing to improving scientific endeavour in this regard.”
  • “I am very enthusiastic about exploiting new ways of sharing my research output beyond the established peer-review journal system.”
  • “I believe that sharing research outputs fully, including data and code are essential to accelerate research, and I have benefitted from it in my own research.”

Summarising, researchers expressed a great desire for contributing to a cultural change. Researchers wanted to change the way in which research was disseminated and to increase research transparency and reproducibility.

Let’s get to work

We all met (the researchers, Wellcome Trust and Cambridge Open Research teams) on Friday 27 January to officially start the two year project. Each research group was appointed a facilitator – a dedicated member of the Cambridge Open Research team to support researchers during the Project. Research groups will meet with their facilitators on a monthly basis in order to discuss shareable research outputs and to decide on best ways to disseminate these outputs. Every six months all project members will meet together to discuss the barriers to sharing discovered and to assess the progress of the Project.

One of the main goals of the Project is to learn what the barriers and incentives are for Open Research and to share these findings with others interested in the subject to inform policy development. Therefore, we will be regularly publishing blog posts on the Unlocking Research blog and on the Wellcome Open Research blog with case studies describing what we have discovered while working together. There will be an update from each research group every six months. We will also be publicly sharing all main outputs of the Project.

We are all extremely excited about going “Open” and we suggest that anyone interested in the Open Research practice watches this space.

Published 08 February 2017
Written by Dr Marta Teperek
Creative Commons License

Creating a research data community

Uncategorized, ,

Are research institutions engaging their researchers with Research Data Management (RDM)? And if so, how are they doing it? In this post, Rosie Higman (@RosieHLib), Research Data Advisor, University of Cambridge, and Hardy Schwamm (@hardyschwamm),  Research Data Manager, Lancaster University explore the work they are doing in their respective institutions.

Whilst funder policies were the initial catalyst for many RDM services at UK universities there are many reasons to engage with RDM, from increased impact to moving towards Open Research as the new normal. And a growing number of researchers are keen to get involved! These reasons also highlight the need for a democratic, researcher-led approach if the behavioural change necessary for RDM is to be achieved. Following initial discussions online and at the Research Data Network event in Cambridge on 6 September, we wanted to find out whether and how others are engaging researchers beyond iterating funder policies.

At both Cambridge and Lancaster we are starting initiatives focused on this, respectively Data Champions and Data Conversations. The Data Champions at Cambridge will act as local experts in RDM, advocating at a departmental level and helping the RDM team to communicate across a fragmented institution. We also hope they will form a community of practice, sharing their expertise in areas such as big data and software preservation. The Lancaster University Data Conversations will provide a forum to researchers from all disciplines to share their data experiences and knowledge. The first event will be on 30 January 2017.

RDMFBreakoutHaving presented our respective plans to the RDM Forum (RDMF16) in Edinburgh on 22nd November we ran breakout sessions where small groups discussed the approaches our and other universities were taking, the results summarised below highlighting different forms that engagement with researchers will take.

Targeting our training

RDM workshops seem to be the most common way research data teams are engaging with researchers, typically targeting postgraduate research students and postdoctoral researchers. A recurrent theme was the need to target workshops for specific disciplinary groups, including several workshops run jointly between institutions where this meant it was possible to get sufficient participants for smaller disciplines. Alongside targeting disciplines some have found inviting academics who have experience of sharing their data to speak at workshops greatly increases engagement.

As well as focusing workshops so they are directly applicable to particular disciplines, several institutions have had success in linking their workshop to a particular tangible output, recognising that researchers are busy and are not interested in a general introduction. Examples of this include workshops around Data Management Plans, and embedding RDM into teaching students how to use databases.

An issue many institutions are having is getting the timing right for their workshops: too early and research students won’t have any data to manage or even be thinking about it; too late and students may have got into bad data management habits. Finding the goldilocks time which is ‘just right’ can be tricky. Two solutions to this problem were proposed: having short online training available before a more in-depth training later on, and having a 1 hour session as part of an induction followed by a 2 hour session 9-18 months into the PhD.

Tailored support

Alongside workshops, the most popular way to get researchers interested in RDM was through individual appointments, so that the conversation can be tailored to their needs, although this obviously presents a problem of scalability when most institutions only have one individual staff member dedicated to RDM.

IMG_20161122_121401There are two solutions to this problem which were mentioned during the breakout session. Firstly, some people are using a ‘train the trainer’ approach to involve other research support staff who are based in departments and already have regular contact with researchers. These people can act as intermediaries and are likely to have a good awareness of the discipline-specific issues which the researchers they support will be interested in.

The other option discussed was holding drop-in sessions within departments, where researchers know the RDM team will be on a regular basis. These have had mixed success at many institutions but seem to work better when paired with a more established service such as the Open Access or Impact team.

What RDM services should we offer?

We started the discussion at the RDM Forum thinking about extending our services beyond sheer compliance in order to create an “RDM community” where data management is part of good research practice and contributes to the Open Research agenda. This is the thinking behind the new initiatives at Cambridge and Lancaster.

However, there were also some critical or sceptical voices at our RDMF16 discussions. How can we promote an RDM community when we struggle to persuade researchers being compliant with institutional and funder policies? All RDM support teams are small and have many other tasks aside from advocacy and training. Some expressed concern that they lack the skills to market our services beyond the traditional methods used by libraries. We need to address and consider these concerns about capacity and skill sets as we attempt to engage researchers beyond compliance.

Summary

It is clear from our discussions that there is a wide variety of RDM-related activities at UK universities which stretch beyond enforcing compliance, but engaging large numbers of researchers is an ongoing concern. We also realised that many RDM professionals are not very good at practising what we preach and sharing our materials, so it’s worth highlighting that training materials can be shared on the RDM training community on Zenodo as long as they have an open license.

Many thanks to the participants at our breakout session at the RDMForum 16, and Angus Whyte for taking notes which allowed us to write this piece. You can follow previous discussions on this topic on Gitter.

Published on 30 November
Written by Rosie Higman and Hardy Schwamm
Creative Commons License