Are journals requiring data sharing mandatory for all research types?

Requirements vary by journal, publisher, and research type. Most journals with data-sharing policies require a Data Availability Statement even when a dataset cannot be openly shared. Clinical-trial data can involve additional journal, funder, ethical, and regulatory expectations. If data cannot be shared, explain the restriction in the statement and follow the journal's current instructions.

Reference notes

Coverage

NIH DMS Policy · 11 journal policy records · 14 repositories

Sources

Feb. 2026 policy snapshot · NIH and Nature Portfolio rechecked Sep. 2026

Last reviewed

Dataset: February 2026 · core sources: September 2026

Prepared by the Manusights editorial team.

Compliance-and-repository guide

Data Sharing Requirements for Biomedical Research

Data sharing in biomedicine went from optional to expected to mandatory over the last decade. The NIH Data Management and Sharing Policy, which took effect January 25, 2023, applies to NIH-funded research that generates scientific data. It requires a data-management and sharing plan, then expects researchers to carry out the approved plan unless a justified limitation applies.

This guide covers what the NIH policy requires, what major journals ask for in data availability statements, where to deposit different types of biomedical data, and what FAIR principles mean in practice.

Quick orientation

Use this page when the manuscript is getting close to submission and the data-sharing plan needs to be concrete, not aspirational.

This guide helps translate policy language into operational choices: what the NIH expects, what journals usually require in the statement itself, which repository fits the data type, and what “FAIR” means at the moment of deposit.

NIH policy summary11 journal-policy records14 repository optionsFAIR in practice

Best used with

Open access guide

Use it when funder policy is also shaping the publication route and APC decision.

Author rights guide

Pair repository planning with the rights and embargo terms that govern manuscript sharing.

Research data management plan

Start there when file organization, storage, security, documentation, roles, and preservation still need to be planned.

The NIH Data Management and Sharing Policy (2023)

The NIH DMS Policy applies to all NIH-funded research that generates scientific data: including grants, contracts, and intramural research. It applies to all applications and proposals submitted on or after January 25, 2023.

The policy requires:

1.Submit a Data Management and Sharing (DMS) Plan with your grant application
2.Data must be deposited in an established repository as soon as possible: no later than publication or end of award
3.Data Management costs are allowable as direct costs (up to ~$30,000/year without special justification)
4.Cite the shared dataset in any resulting publications

What "scientific data" means under the policy: Recorded factual material commonly accepted in the scientific community as necessary to validate research findings, not lab notebooks, preliminary analyses, or materials that would typically be in a methods section.

Exceptions: Data may be exempt from sharing if sharing is restricted by law (HIPAA, tribal law), contract, or if the data contains information that can't be de-identified. Document the reason in your DMS Plan.

NIH Institute-specific policies: Some NIH institutes have stricter requirements on top of the base policy: NHGRI for genomics, NCI for cancer data. Check your funding institute's data sharing policy in addition to the base DMS Policy.

Other Major Funder Data Sharing Mandates

Wellcome Trust

Since 2017 (updated 2021)

Strong data sharing mandate. Requires a data management plan, data sharing in an appropriate repository, and a data availability statement in all publications. All publications must be OA (CC BY), and the underlying data must be available.

UKRI (all councils)

Since 2015 (RCUK policy, now UKRI)

Research outputs including data must be made available as openly as possible. Data underlying publications must be deposited in an appropriate repository. Applies to BBSRC, MRC, ESRC, EPSRC, and other UKRI councils.

Gates Foundation

Since 2017

One of the strictest. It requires CC BY for all publications and immediate open access to underlying data. Data sharing plan required with grant application.

European Research Council (ERC)

Since 2017 (expanded Horizon Europe)

Open Research Data pilot is now the default for all funded projects. Requires a Data Management Plan and deposition in a trusted repository where possible.

Journal data-availability policies

Compare how strict major journal families are about data availability statements and repository deposition. Export the current view or copy rows into a lab or library data-sharing guide.

Journal group	Strictness	Policy summary
Nature family	Strict	Mandatory data availability statement. Raw data for figures required. Specific repositories required for genomics, structures, sequences. Code must be deposited.
Science / AAAS	Strict	Mandatory data availability statement. All data must be available to reviewers and readers. Structured data deposited in appropriate repositories.
Cell Press	Strict	Data availability statement required. Original data for all figures must be deposited or available on request. Specific NCBI/PDB repositories for applicable data types.
NEJM	Moderate	Data availability statement required. For clinical trials, data sharing plan must be registered. Patient-level trial data sharing increasingly expected.
JAMA Network	Moderate	Data and statistical code availability statement required. Deidentified participant data must be available for clinical trials with planned sharing.
Lancet family	Moderate	Data sharing statement required. Original data must be available on reasonable request. Clinical trial data sharing plan required.
BMJ	Moderate	Data availability statement required. Supports open data; where possible, data should be deposited. Structured data sharing encouraged via OSF or similar.
PLOS ONE / PLOS Medicine	Strict	All data underlying figures and results must be fully available. Data deposited in appropriate repository or included as supplementary material. No 'available on request': must be actually available.
eLife	Strict	All data must be available. No 'available on request.' Code must be available. Extremely transparent data sharing expectations.
PNAS	Moderate	Data availability statement required. Data must be deposited in an appropriate repository where one exists.
Nature Communications	Strict	Same as Nature family: data availability statement, repository deposition for applicable data types.

Where to deposit your data

Search repositories by data type, cost, or notes. Export the current view or copy rows directly into a DMS plan, lab SOP, or author-support page.

Repository	Data type	Cost	Notes	URL
NCBI GEO	Gene expression, genomics arrays	Free	Required by most molecular biology journals for microarray and RNA-seq data. Generates GSE accession number.	www.ncbi.nlm.nih.gov/geo/
NCBI SRA (Sequence Read Archive)	Raw sequencing data (NGS, WGS, RNA-seq)	Free	Required for raw sequencing data. Submit before manuscript submission to include accession in paper.	www.ncbi.nlm.nih.gov/sra
NCBI dbGaP	Human genomic + phenotype data	Free	NIH-approved repository for controlled-access human genomic data with identifiability concerns.	www.ncbi.nlm.nih.gov/gap/
NCBI dbSNP / ClinVar	Genetic variants	Free	For novel SNPs and variant-disease associations.	www.ncbi.nlm.nih.gov/snp/
Protein Data Bank (PDB)	Protein and nucleic acid structures	Free	Required by Nature, Cell, and structural biology journals for all new structures. Generates PDB ID.	www.rcsb.org
Zenodo	Any type (general purpose)	Free up to 50GB	CERN-hosted, FAIR-compliant, generates DOI. Good for computational code, datasets, and materials not accepted elsewhere.	zenodo.org
Figshare	Any type (general purpose)	Free up to 20GB	Good for figures, datasets, posters, and code. Many institutions have sponsored accounts. Generates DOI.	figshare.com
OSF (Open Science Framework)	Any type (general purpose)	Free (5GB private/project)	Good for pre-registration, protocols, and data. 5GB private storage per project; public projects have higher limits. Strong in psychology/social science; growing in biomedicine.	osf.io
Dryad	Any type (general purpose)	$120 DPC (often journal-covered)	Partner with many journals: some journals automatically transfer data to Dryad at acceptance.	datadryad.org
GitHub / Zenodo (code)	Analysis code and software	Free (code); Zenodo for archival DOI	GitHub for active development; Zenodo integration for a frozen, citable version. Required for computational methods papers.	github.com
PhysioNet	Physiological and clinical signals	Free	For ECG, EEG, ICU monitoring data. Strong in clinical and biomedical signals.	physionet.org
ImmPort	Immunology data	Free	NIH-supported repository for immunological data including flow cytometry and clinical trial immunology data.	immport.org
The Cancer Imaging Archive (TCIA)	Medical imaging (CT, MRI, PET)	Free	For de-identified cancer imaging datasets. Required/recommended by imaging and oncology journals.	www.cancerimagingarchive.net
Harvard Dataverse	Any type (general purpose)	Free	Well-regarded general repository. Many universities have institutional Dataverse installations.	dataverse.harvard.edu

FAIR Principles in Practice

FAIR data principles (Findable, Accessible, Interoperable, Reusable) are now the standard framework for data sharing. NIH, Wellcome, UKRI, and most major journals reference FAIR compliance as the goal. Here's what each means operationally:

Findable

Deposit in a repository that assigns a persistent identifier (DOI or accession number). Include rich metadata so search engines can index it. Don't just put data on a lab website that disappears.

Accessible

Data should be retrievable via open protocols. For controlled-access data (e.g., dbGaP), the access procedure itself must be publicly documented. The process for requesting access counts as 'accessible' even if the data aren't open.

Interoperable

Use standard file formats (CSV over Excel, FASTQ over proprietary formats, TSV over custom delimiters). Include a data dictionary. Make it so someone without your specific software can use the data.

Reusable

Attach a license (CC BY for open data). Include enough provenance (how the data were collected, processed, and what quality checks were applied) so that someone else can reproduce your decisions.

Practical note

Three data-sharing mistakes that create late-stage submission problems

Waiting until the manuscript is drafted to identify the repository, accession timing, or data-availability wording the journal expects.

Assuming a general-purpose repository is always acceptable when a domain-specific archive is the norm for that data type.

Treating FAIR as a slogan rather than a deposit checklist covering identifier, metadata, format, and reuse information.

Most data-sharing problems surface late, when the manuscript is already in editorial review. Use the data and availability checks in the manuscript submission checklist to catch a missing repository, accession, or FAIR gap before an editor does.

References

National Institutes of Health. (2023). NIH Data Management and Sharing Policy. U.S. Department of Health and Human Services. [sharing.nih.gov ↗]
Nature Portfolio. (2026). Reporting standards and availability of data, materials, code and protocols. Rechecked September 2026. [nature.com ↗]
Wilkinson MD, Dumontier M, Aalbersberg IJ, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016;3:160018. [doi.org/10.1038/sdata.2016.18 ↗]
UK Research and Innovation (UKRI). Research data management and sharing policy. Retrieved February 2026. [ukri.org ↗]
Zenodo. About Zenodo: open repository for research. CERN. Retrieved February 2026. [zenodo.org ↗]
Figshare. About figshare: open data repository. Digital Science. Retrieved February 2026. [figshare.com ↗]
Dryad. About Dryad: open data repository. Dryad Digital Repository. Retrieved February 2026. [datadryad.org ↗]

Data note: The repository and journal-policy rows are a February 2026 planning snapshot; the NIH policy and Nature Portfolio source were rechecked in September 2026. Funder, journal, repository, and data-access conditions evolve, so always check your funder's current policy and the journal's author instructions before submission. For accuracy corrections or updates, contact hello@manusights.com.

Cite this resource

Citation formats

Free to use and cite, with attribution. See the reuse policy for what you may republish.

APA: Manusights. (2026). Data sharing requirements for biomedical research: NIH policy, journals, and repositories (Dataset snapshot February 2026; NIH and Nature Portfolio sources rechecked September 2026). https://manusights.com/resources/data-sharing-requirements
MLA: Manusights. "Data Sharing Requirements for Biomedical Research: NIH Policy, Journals, and Repositories." Manusights, dataset snapshot Feb. 2026; NIH and Nature Portfolio sources rechecked Sept. 2026, manusights.com/resources/data-sharing-requirements.
Vancouver: Manusights. Data sharing requirements for biomedical research: NIH policy, journals, and repositories [Internet]. Dataset snapshot Feb 2026; NIH and Nature Portfolio sources rechecked Sep 2026. Available from: https://manusights.com/resources/data-sharing-requirements

Related guides in this collection

Open Access Guide

Use this when repository and data-sharing decisions also depend on funder OA mandates or APC planning.

Author Rights Guide

Pair data-sharing planning with the rights and embargo rules that govern how manuscripts and accepted versions can be shared.

Research Data Management Plan

Plan data organization, documentation, storage, security, access, preservation, and team responsibilities before sharing.

← Preprint Guide All Resources Reporting Guidelines →

Frequently Asked Questions

Are data sharing requirements mandatory for all research types?

Requirements vary by journal, publisher, and research type. Nature Portfolio journals, Cell Press, PLOS journals, and BMJ require data sharing for all original research by default, with exceptions for privacy-restricted data. Most journals with data sharing policies require a Data Availability Statement in the manuscript regardless of whether data can be fully shared. If data cannot be shared, you must explain why - privacy, legal restrictions, or third-party ownership are all accepted reasons at most journals.

Where should I deposit research data before submitting a manuscript?

The right repository depends on your data type. Genomics data goes to NCBI (GEO for gene expression, SRA for raw sequencing, dbGaP for controlled human genomics data). Protein structures go to the RCSB Protein Data Bank. Clinical trial data goes to ClinicalTrials.gov. For general research data without a domain-specific repository, Zenodo, Figshare, or Dryad are widely accepted. Most journals provide a list of recommended repositories in their data sharing policy.

What should a Data Availability Statement include?

A Data Availability Statement should specify: (1) where the data can be accessed (repository name and URL or DOI), (2) any access restrictions and the reason for them, and (3) the accession number or identifier for the deposited dataset. If all data are in the manuscript itself (in figures and tables), state that explicitly. If data cannot be shared due to privacy or ethical restrictions, state the restriction and whether data can be requested from the corresponding author. Many journals provide a template - use it, since reviewers and editors check the statement during review.