If a dataset in OSF leads to a publication, the data need to be deposited in a repository to provide a unique and persistent reference to the data (this is not possible in OSF, because users can edit, add, remove files, even if there is a DOI at project level). DataverseNL would be a logical choice (connection with OSF is also possible there), but external repositories or Yoda would be possible as well. We need to write a page about this. This could be a guide, or it could be incorporated in the OSF manuals (to be decided later).