Data Sharing#

Background#

Data is key to neuroscience, and it should be shared for maximal utility. There is increasing pressure from funders and publishers to share data alongside publications:

Wellcome:

“We expect our researchers to maximise the availability of research data, software and materials with as few restrictions as possible. As a minimum, the data underpinning research papers should be made available to other researchers at the time of publication, as well as any original software that is required to view datasets or to replicate analyses.”

Cell Press:

“All datasets that are composed of standardized datatypes and reported in Cell Press papers must be deposited in a datatype-specific, Cell-Press-recommended repository before a paper is accepted for publication.”

“All original code must be either deposited in a repository that mints DOIs or included in the supplemental information before a paper is accepted for publication.”

Requirements#

Currently, SWC does not have any specific code and data sharing requirements, but your funder or publisher may do.

Ideally, all data and code will be shared in such a way that any other researcher could easily reproduce key figures from a publication, and build on those data and software in their own research.

Code#

Original code can be hosted in a public GitHub/GitLab repository, hosted on an organisational account (i.e. not an individual account). This could be the SWC organisation, or that of a particular team, lab or collaboration. To enable reuse, code should include an appropriate license.

To ensure a permanent archive of the code associated with a publication a DOI should be created (e.g. using Zenodo).

Data#

Repository#

There are many repositories that will accept the types of data generated by SWC. They come into three categories, general, neuroscience-specific and modality-specific. If you do not have a preference, we recommended:

For small (<20GB) datasets: Use the existing SWC section of the UCL FigShare instance. This is easy to use, and is supported centrally by UCL

For larger datasets: Create a GIN repository, ideally within the SWC organisation. GIN uses Git and provides a versioning system, making it particularly well suited for ongoing research projects.

Specific modalities: In some cases, there may be a modality or field-specific repository that would be best suited for the data (or mandated by the publisher), some examples include:

Brain Image Library - Large brain image datasets
DANDI - Electrophysiology, optophysiology, behavioural time-series (must be in NWB format)
BioImage Archive - Biological images

In all cases, using a repository that can create a DOI will help increase citations.

Structure#

There is no single best way to structure your data, but it must be in a form that allows users to easily reuse it for their own purposes. This may be by adopting specific file formats (such as NeuroData Without Borders). Alternatively the data could be organised within a standardised project folder structure. Many labs have their own system, but those starting from scratch may wish to use the NeuroBlueprint data structure.

However the data is structured, all relevant metadata must be included with the raw data.

Assistance#

If you have any technical questions about sharing data or code, please contact adam.tyson@ucl.ac.uk.

Reference of potential repositories for neuroscience data#

Repository	URL	Domain
Globus		Anything
figshare	https://rdr.ucl.ac.uk/The_Sainsbury_Wellcome_Centre	Anything
Zenodo	https://zenodo.org/	Anything
Dryad	https://datadryad.org	Anything
GIN (German Neuroinformatics Node)	https://gin.g-node.org/	Neuroscience
EBRAINS	https://ebrains.eu/service/share-data	Neuroscience
Open Source Brain	https://www.v2.opensourcebrain.org/	Any neuroscience data (data must be hosted elsewhere)
Brain Image Library	http://www.brainimagelibrary.org/	Large brain image datasets
Image data resource	https://idr.openmicroscopy.org/	Reference image datasets
BioImage Archive	https://www.ebi.ac.uk/bioimage-archive/	Biological images
DANDI	https://dandiarchive.org/	Electrophysiology, optophysiology, behavioural time-series and images from immunostaining experiments
NeMO	https://nemoarchive.org/	Omic data from the BRAIN Initiative (& others)
Openneuro	https://openneuro.org/	BIDS-compliant MRI, PET, EEG etc
CRCNS	https://crcns.org/	Computational neuroscience
BrainGlobe	https://gin.g-node.org/BrainGlobe/atlases	Brain Atlases
NeuroMorpho	https://neuromorpho.org/	Neuronal morphologies
Cell Image Library	http://ccdb.ucsd.edu/home	Cell images
ModelDB	https://modeldb.science/	Computational neuroscience models