blog

Recently, we published a slide set that was presented during the International Zika Summit held at the Institut Pasteur, Paris on April 25-26 2016 which explores collaborative data sharing in outbreak situations. We reached out to the authors and are delighted to have them write a guest post for us to discuss a possible way to share and collaborate with data effectively in times of crises.

This post is written jointly by Antonio Borderia and Magnus Fontes (International Network for Data Analysis, Institut Pasteur Paris, France), Cameron MacPherson and Stephanie Thomas (Milieu Intérieur, Institut Pasteur Paris, France), Lara Mangravite and Larsson Omberg (Sage Bionetworks, Seattle, USA).

The recent Zika outbreak has seen many different institutes and organizations join forces to facilitate data sharing as an essential step towards improving the health crisis as fast as possible (i.e: Wellcome trust data sharing policy, The global health network website on Zika, and the WHO about data sharing during public health emergencies). Scientific journals have also stepped forward allowing for early release of results and data without repercussions on subsequent publications (Nature Microbiology, New England Journal of Medicine, F1000Research, PLoS, and others). While these important actions target several barriers hindering data sharing they do not provide the online tools required to track and manage ownership of data or the sharing of other digital assets beside data. Indeed, reducing the time to action in outbreak situations would be best facilitated through a globally coordinated effort for joint analysis on shared data. For such efforts to succeed in the context of the current scientific environment, it is paramount that individuals receive attribution for their contributions of results and ideas as well as of data. If done correctly, trust becomes implicit.

Time is a fundamental factor during outbreak situations but is something that is exasperated by the legal, ethical, and communication hurdles that are present in large international collaborations.  The reduction of technical barriers for sharing such as file sharing tools like Dropbox is also a major source of concern for issues surrounding security and data governance. Recognising this, scientific journals are establishing best-practices for data sharing (See “Rebooting review” in Nature Biotechnology) and several organisations are developing online software solutions that mitigate the security risks implicit in collaborative research (See software references at the bottom of the post). Of the several platforms available for data sharing, the Synapse platform (see Box 1) allows for tracking of data provenance while maintaining strong restrictions on data sharing and governance. Synapse was developed by Sage Bionetworks to enable collaboration and sharing during the research process instead of the traditional model of sharing only upon publication.  This development has been undertaken in collaboration with several large research consortia, such as The Cancer Genome Atlas, DREAM Challenge initiative and accelerating Medicine Partnership for Alzheimer’s Disease. We have adopted Synapse for all of our collaborative projects as it allows fine-grained control over user access and data-specific ethical compliance. Here, we announce a model for collaborative science in outbreak situations by providing a template Synapse project for the Zika crisis, “ZikaNet”. All interested parties are welcome to join the project, to contribute together, and to share their data in full confidence, knowing that ownership of the data is automatically tracked by Synapse. 

In the ZikaNet project, each laboratory or group working on Zika is provided their  own private space in which to upload data and analysis. Within their private space, each lab can enforce their individual needs of security and compliance – choosing to share all, or part of, their content with other labs. The ZikaNet project also provides a common workspace where groups can share resources and work together. The resource is designed to bring contributing scientists front and center, by ensuring attribution for individual contributions throughout the span of the analysis.  Researchers are able to track the “impact” of their shared work as it is integrated into downstream analyses and to mint a DOI (digital object identifier) unique to each result and/or data file. DOI not only facilitates publication, but it also provides a digital mechanism for anyone with the published DOI to access the full provenance of the data and subsequent insights. Mainstreaming the use of DOI could also help to steer journals policies to allow for early release of data and results. This system is designed to reduce barriers to collaborative research typically encountered in highly competitive fields through the facilitation of data sharing, provision of transparency and attribution in research findings and the advancement of science. We envision that this approach will prove valuable in the very challenging situations of an outbreak, where coordination and trust is key to acting rapidly.

Software references

Transmart (https://transmartfoundation.org/)

Open Science Framework (https://osf.io/)

Globus (https://www.globus.org/)

Labkey (https://www.labkey.com/)

Dropbox (https://www.dropbox.com)

Google Drive (https://www.google.com/drive/)