Chapter 6: Data Sharing and Access

6.1 The Case for Data Sharing

Data sharing accelerates scientific discovery, enables validation and reproducibility, and maximizes return on research investment. Shared data enables meta-analyses with greater statistical power, allows replication of findings, facilitates new discoveries from existing datasets, and supports training of future researchers.

Major funders, publishers, and research institutions increasingly mandate or incentivize data sharing. The question is no longer whether to share, but how to share effectively while protecting participant privacy and researcher interests.

6.2 Research Data Repositories

Repositories provide infrastructure for data preservation, discovery, and access:

Repository Types

Type	Examples	Best For
Domain-Specific	dbGaP, EGA, ArrayExpress, PRIDE	Data with established community standards
Generalist	Zenodo, Figshare, Dryad, OSF	Data without domain repository
Institutional	University data archives	Institutional compliance, local support
Controlled Access	dbGaP, EGA, YODA, Vivli	Individual-level sensitive data
National	UK Data Service, ICPSR	Large surveys, government data

6.3 Data Deposit Process

Preparing data for repository deposit requires careful attention to documentation and formatting:

Deposit Preparation

• Data Preparation: Clean, well-organized files in open formats
• Documentation: README, data dictionary, methodology
• Metadata: Complete repository metadata forms
• Access Level: Determine open vs. controlled access
• License Selection: Choose appropriate data license
• Linking: Connect to related publications, code, protocols

Data Licensing

License	Permissions	Conditions
CC0	Maximum reuse, public domain	No conditions (waive all rights)
CC-BY	Reuse with attribution	Credit original creators
CC-BY-NC	Non-commercial reuse	Attribution, non-commercial only
DUA	Custom terms	Varies by agreement
Open Data Commons	Database-specific	PDDL, ODC-By, ODbL

6.4 Clinical Trial Data Sharing

Clinical trial data sharing presents unique considerations:

Trial Data Sharing Platforms

Platform	Data Source	Access Model
ClinicalStudyDataRequest	Industry sponsors	Proposal review, controlled
Vivli	Industry, academic trials	Proposal review, virtual enclave
YODA Project	Johnson & Johnson trials	Proposal review, data transfer
NHLBI BioLINCC	NHLBI-funded trials	Request and agreement
IMMPORT	Immunology trials	Open or controlled

6.5 Collaborative Data Networks

Federated approaches enable multi-site research without centralizing data:

Network Models

• OHDSI: Observational health data science using OMOP CDM across 700+ institutions
• PCORnet: Patient-Centered Outcomes Research Network with common data model
• TriNetX: Global health research network with federated queries
• EHDEN: European Health Data Evidence Network using OMOP
• GA4GH Beacon: Federated discovery of genomic variants

6.6 Data Citation

Proper citation gives credit to data creators and enables tracking of data reuse:

Data Citation Principles

• Importance: Data citations are legitimate and important research contributions
• Credit: Citation should facilitate attribution and credit
• Evidence: Citations should support claims made about data
• Identification: Citations should include persistent identifiers
• Access: Citations should facilitate access to data and metadata
• Persistence: Citations should be as durable as cited objects

Citation Format

Creator (Publication Year). Title. Version. Publisher/Repository. Identifier

Example: Smith J, Jones M. (2024). Diabetes Biomarker Study Dataset. V2. Zenodo. https://doi.org/10.5281/zenodo.123456