Interview: Patrik Magnusson about working for Open Science at KI

What is your approach to open science?

– An important entry point to open science for me has been that the field which I primarily work in, genetic epidemiology, has undergone a dramatic change. We have moved from a situation with a lot of publication bias and a high proportion of false positive findings, to being one of the fields that is doing best in terms of replicability and transparency. The reason we know this is because we have moved from looking at single genetic variants to measuring variation across the whole genome and have scaled up studies accordingly. We have a thousand times larger studies today than we did before and that means we have the statistical power to analyse the whole genome. Researchers have shared their results and together they have been able to do meta-analyses.

– For me, it's been a real eye-opener and a fantastic development. I've been really inspired by the developments that have taken place in genetic epidemiology and I hope it will lead to other fields also adopting similar practices tackling the publication bias, such as the practice of systematically publishing aggregated results.

What is the Open Science Working Group's mission?

– We are a preparatory working group, so we have no decision-making power on our own, but we will be producing documents for the Committee for Research to consider. They will then give us guidance about which direction we should work towards. Since open science is so multifaceted, part of the task is to specify how KI should define the concept and what it should consist of. It’s a bit of definition work.

– One concrete task is that we will develop an action plan for the promotion and incorporation of an open science system by 2026. By then, per government decree, higher education institutions in Sweden are obliged to have an open science system in place. As I see it, quite a lot of preparatory work, analysis and anchoring work is required. De facto, we have a scientific system based on competition, where researchers compete to publish results that advance the field, and lead to innovation and development. There is some protectionism built into the system. So, it's easier said than done.

What have you done so far?

– We've had a handful of meetings and it's almost time to report back to the Committee for Research again. A lot of the meetings that we've had have really been about educating each other in the group based on our respective horizons, knowledge and expertise. It's been interesting just learning from each other.

– We've also started to decide how we're going to structure the work of developing our action plan, as well as started a search for KI’s policies, fishing out what policies there are and what they sound like in today's open science climate. It's clear that some need to be revised.

– We are also members of various collaborative forums. SUHF has a reference group, the Swedish Research Council has another one and so does European Open Science Cloud (EOSC).

What else can you tell us about the action plan you will develop?

– Regarding the action plan that KI should develop and agree on, there is a document that SUHF has developed. It consists of a roadmap with recommendations to the different universities. We have studied it and decided that we will use it as a guide.

Is there an idea to harmonise the policies that you choose to implement here at KI with those that apply nationally and internationally?

– Yes, as much as possible. One typical thing that I've been struck by and that I think is complicated is that a lot that has to do with openness is based on the condition that there are no legal reasons to be closed. Maybe half of the research at KI is based on sensitive personal data that needs to be protected for privacy reasons. Then there are a lot of studies that have potential commercial beneficiaries, where there are agreements with companies and patents. So it's quite difficult to harmonise everything at once with intellectual property rights, teachers' exemption and so on. It becomes an interpretation of law all the time, of how important is it to do what we are supposed to do, which is to conduct research in the best and most transparent way possible, versus how big is the risk that we disclose sensitive data or make commercialization impossible. Taking as a starting point recommendations made by SUHF and research funders for transparency requirements, I think the implementations will eventually be quite similar. The FAIR principles of open science are central and widely accepted.

You have mentioned some challenges with open science. Where do you see benefits and opportunities?

– The benefits are that a transition to an open science society leads to greater transparency and more reliable research. It means lower publication bias and lower rates of false positive findings, which are really the same thing. It will be possible to recapitulate and validate each other's data and findings. I think research is highly regarded because it is undeniably the best we can do to achieve new knowledge. It is not perfect, but it is the best we can do. Still, there is a lot of room for improvement in research methodology. If research is to continue to be the thing to hold onto when the wind blows and guidance towards the truth is needed, then it is important that we do as well as we possibly can. This is an important process in that direction.

– In terms of opportunities, I think it can lead to more efficient and cost-effective research, because when people start sharing data, then you can use, reuse and co-analyze the data and get better and more accurate estimates. Hopefully, research results will also be published in a more cost-effective way.

– I also think this could lead to less bias, less nepotism and a fairer system.

Is there anything else you would like to say to KI?

– Make research more efficient, for example by making data sharing easier and possible. Use the tools and structures that already exist. Facilitate and enable meta-analysis of many studies. Regardless of everything else about open science, this is stuff that is desirable and will lead to better and truer results. I think we can take that with us. Some of the changes will be met with friction, but I think it's the right way to go.

– Then research funders and research evaluators will have to find factors that reward us as researchers when we make our data FAIR, and are generous with our knowledge, methods, data and results. So it shouldn't be impossible, but it needs to be valued highly, more than publishing in high impact factor journals.