Open access to research data makes use of resources more efficient

Representatives of data sharing programs and of public universities in São Paulo State attend workshop at FAPESP to report on experiences and challenges of building a culture of open data and open science.

By André Julião  |  Agência FAPESP – Nine months after Brazil was reported to be the country with the largest number of open-access scientific articles in the world, national research funding agencies from 11 European countries joined the European Commission (EC) in announcing that by 2020, all recipients of public research funding will be required to publish only in open-access journals under a license enabling anyone to reuse the material free of charge.

The measure appears to be an inevitable choice at a time when open-access publication of research is growing, and there are more structures and policies designed to make even data not associated with publications available to other researchers.

“The goals in both the Netherlands and Europe are very ambitious. We see open science as the future. For this reason, we’re proposing policies and building infrastructures to make data sharing a reality,” said Ingrid Dillo, Deputy Director of Data Archiving and Networked Services (DANS), the Dutch national institute for permanent access to digital research resources. Funded by the Royal Netherlands Academy of Arts and Sciences (KNAW) and the Netherlands Organization for Scientific Research (NWO), for ten years, DANS has supported the efforts of Dutch researchers to preserve and share data over the long term.

Dillo was one of several foreign guest speakers to address Open Science/Open Data: Challenges and Best Practices in Scientific Data Management, a workshop held in September 2018 at FAPESP’s headquarters in São Paulo, Brazil.

The event was also attended by representatives of São Paulo State’s seven public universities, who presented their data sharing and management initiatives. These universities are part of a working group coordinated by FAPESP to create an open-access network of research data repositories. FAPESP recognizes the importance of appropriate data management as an essential part of best practice in research.

The event also featured a presentation on the activities of SciELO (Scientific Electronic Library Online, a project supported by FAPESP) with regard to open-access publications, a key component of open science.

“We discussed the new practices and policies implemented by FAPESP to promote open science and open data,” said the workshop’s organizer, Claudia Bauzer Medeiros, a professor in the University of Campinas’s Institute of Computing (IC-UNICAMP) and a member of the steering committee of the FAPESP Research Program on eScience and Data Science.

“In the Netherlands, they’ve been working on the development of an open data system for ten years, so this is consolidated now. In Canada, it’s a more recent initiative, but they already have large repositories that could help us in Brazil,” Medeiros told Agência FAPESP.

“What we’re trying to do in Canada is create partnerships between various sectors, such as libraries, researchers and infrastructure providers,” said another speaker at the event, Lee Wilson, Systems Manager of Portage Network, an initiative launched in 2015 by the Canadian Association of Research Libraries (CARL). “These partnerships merely reinforce the efforts of each group to enable us to create better services for our researchers. This can be a model for Brazilian funding agencies.”

Open data: open science

“The idea of open data is that many researchers can collaborate both by doing research together and by sharing knowledge or reusing each other’s results, data and methods. This helps create technology, science, knowledge, and economic and social wealth,” Medeiros said.

An example of data sharing can be found in genomics studies, such as the research conducted by Brazilian Benilton de Sá Carvalho, a researcher at the University of Campinas’s Mathematics, Statistics and Scientific Computation Institute (IMECC-UNICAMP) and a member of the steering committee of the Brazilian Initiative on Precision Medicine (BIPMed).

Carvalho is one of the scientists responsible for the construction of a new tool that could facilitate the identification of disease-causing genetic variants (read more at agencia.fapesp.br/25254).

“The cost of producing data in the field of genomics is relatively high, and the availability of individuals who are willing to donate DNA and RNA samples is limited, so open data and open science have been key elements of our strategy,” Carvalho said.

He explained that he uses data produced by other research groups to boost the statistical power of the inferences he makes in his analysis, for example. “Two groups can work together on an analysis of the data created by each group separately,” said Carvalho, who is also a staffer in UNICAMP’s Research Pro-Rectorate and represented the Pro-Rector at the workshop.

Transparency

In addition to the sharing of data and results, which can be reused in other research projects, the advantages of open science include transparency and reproducibility. It is no accident that the Netherlands has such a strong initiative in this field. Dutch cases of data fraud that came to light in 2011-12 led to resignations and a major reputational crisis for the universities that employed the researchers who had forged results.

“The data fraud cases in the Netherlands proved to be a blessing in disguise [because they helped us get support to make sure no such cases ever happen again]. A social psychologist and people in medical sciences falsified results. They produced articles that drew a great deal of attention, but eventually, it was discovered that they hadn’t done genuine research or the data didn’t exist,” Dillo said.

The crisis was a huge blow for Dutch universities. “The silver lining is that the universities are now led by people who pay much more attention to the importance of good data management. After these cases, they went to considerable lengths to implement data policies and to create computational infrastructures for proper data storage,” she added.

The presentations delivered to the workshop on Open Science/Open Data: Challenges and Best Practices in Scientific Data Management are available at www.fapesp.br/eventos/opendata.