Data sharing and availability

Findings that cannot be checked are of limited use. We ask authors to make the data behind an article available wherever it is lawful and ethical to do so, following the FAIR principles — findable, accessible, interoperable, reusable — and the ICMJE requirements for clinical trials.

Data availability statement

Every research article must include a data availability statement, placed before the references and published with the article. It must say where the data are, how to get them, and under what conditions. Acceptable forms include:

• The data are openly available in a named repository, with the accession number, DOI or a permanent link.
• The data are included in the article or its supplementary material.
• The data are available from the corresponding author on reasonable request. This is the weakest form and we discourage it; where it is used, the corresponding author is expected to honour requests, and failure to do so is a matter the editor may act on.
• The data are subject to restrictions — participant confidentiality, third-party ownership, legal constraint — in which case state the restriction, who controls access, and the process for applying.
• No new data were generated.

A statement that simply says “data are available on request” without naming who to ask, or “not applicable” for a study that plainly generated data, will be returned before review.

Repositories

Deposit in a recognised subject repository where one exists for your field, or in a general repository such as Zenodo, Dryad, figshare or an institutional repository. Deposit sequence data, structures, proteomics, and other data types in the community-standard archive for that type. Cite the dataset in the reference list with its DOI, so it is discoverable and citable in its own right.

Where participant privacy prevents open deposit, use a controlled-access repository rather than declining to deposit at all.

Code and software

Where results depend on custom code, deposit it in a repository with a DOI and state the licence. Give the version, environment and any parameters needed to reproduce the analysis. Commercial software should be named with its version.

Clinical trials

In line with ICMJE, reports of clinical trials must include a data sharing statement covering: whether individual participant data will be shared; which data; whether supporting documents such as the protocol, statistical analysis plan and informed consent form will be available; when and for how long; with whom; for what types of analysis; and by what mechanism. Where data will not be shared, the statement must say so plainly.

The registered protocol and statistical analysis plan should be submitted with the manuscript.

Materials

Authors are expected to make unique materials — cell lines, plasmids, strains, antibodies, reagents — available to qualified researchers on reasonable terms, and to state any material transfer agreement that applies.

Reporting on data

Report enough to allow the analysis to be repeated: sample sizes and how they were determined, inclusion and exclusion criteria, how missing data were handled, the statistical tests with their assumptions, effect sizes with confidence intervals, and the software used. Do not report a p-value alone.

Editors may request the data

Editors may request underlying data, original unprocessed images or analysis code at any point during review or after publication, where integrity is in question. Refusal without an adequate reason is grounds for rejection, or for an expression of concern or retraction if the article is already published.

Preprints and prior dissemination

Depositing data or posting a preprint before submission does not count as prior publication. Declare both on submission.

Retention

Authors should retain raw data for at least the period required by their institution or funder, and in any case for a period sufficient to answer questions after publication.