Skip to content

Investigate DataCite DOI prefix for NEMAR via UCSD #99

Closed

Description

@neuromechanist

Summary

NEMAR currently uses Zenodo for DOI minting, which means all DOIs use Zenodo's prefix (10.5281). To have professional, NEMAR-branded DOIs (e.g., 10.XXXXX/nemar.nm000103.v1.0.0), we need our own DataCite DOI prefix.

OpenNeuro achieves this by operating under Stanford University Libraries' DataCite membership (sul.openneuro, prefix 10.18112). NEMAR can follow the same pattern under UCSD's DataCite consortium membership.

Background

NEMAR previously relied on Stanford's OpenNeuro for dataset hosting and DOI assignment. We are now transitioning to self-hosted infrastructure at UCSD with:

  • Version-controlled storage on S3 with object locking
  • Dataset mirroring on SDSC (San Diego Supercomputer Center)
  • Full CLI and backend for dataset management and publishing

Action Items

  • Research DOI prefix options and DataCite membership models
  • Identify UCSD contact: Research-Data-Curation@ucsd.edu
  • Confirm UCSD is part of a DataCite consortium
  • Send inquiry email to UCSD Research Data Curation team
  • Confirm they can create a repository account and prefix for NEMAR
  • Obtain DOI prefix assignment
  • Implement DataCite REST API client (supplement or replace Zenodo integration)
  • Design DOI suffix structure: nemar.{datasetId}.v{version}
  • Implement concept DOI support with IsVersionOf/HasVersion relations

Context

Benefits of Own Prefix

  • Professional DOIs: 10.XXXXX/nemar.nm000103.v1.0.0 instead of 10.5281/zenodo.12345678
  • Full control over DOI structure and metadata
  • Proper concept DOI support (something OpenNeuro lacks)
  • Consistent with how major data repositories operate

Activity

  1. neuromechanist commented on Feb 3, 2026

    @neuromechanist
    ContributorAuthor
  2. arnodelorme commented on Feb 3, 2026

    @arnodelorme
  3. neuromechanist commented on Feb 4, 2026

    @neuromechanist
    ContributorAuthor

    Update: UCSD Response (2026-02-04)

    UCSD Library responded positively. Key findings:

    • UCSD uses EZID (operated by California Digital Library / CDL) for DOI minting
    • EZID provides DataCite DOIs to UC affiliates
    • No user fees or per-DOI charges
    • Meeting scheduled for Thu Feb 5 at 11am to discuss workflow and account setup

    Open Question

    Need to confirm whether EZID allows custom-structured suffixes (e.g., nemar.nm000103.v1.0.0) or only auto-generated opaque suffixes. This is the key differentiator vs staying with Zenodo.

    EZID Concepts

    • EZID uses "shoulders" (sub-prefixes within a DOI prefix)
    • Supports both auto-generated ("minting") and custom suffixes ("creating")
    • EZID docs discourage brand strings in suffixes but technically allow them
    • Reference: https://ezid.cdlib.org/learn/id_basics
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    featureNew feature or enhancement

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions