NIH provides guidance on repositories for data sharing and on how to select a data repository. Additional options that provide a fuller range of options may be found by consulting a global list of repositories in biology, medicine and omics. 

Check to see if your funder or journal has repository preferences or publishing requirements for specific data types. Also consult your institutional library for information on institutional repositories and data retention policies.

Most funders like NIH and journals require sharing of data underlying a published article.  However, researchers are often strongly encouraged to share the overall set of data collected in the final version of a research project. Doing so can maximize the impact of your research, as it then can be used to answer an even broader set of research questions.   

Additionally, it is valuable, and often required, to share related outputs from the research, specifically any software created and the code used to analyze the data and create the results. It is generally valuable to publish code along with the data in a repository, but other options exist. For additional resources and guidance, see the NIH Best Practices for Sharing Research Software.

Consult with your Institutional Review Board or related ethics committee to ensure alignment with your data sharing plans. If you are working with genomic data, see the NIH Genomic Data Sharing Policy and guide to Using Genomic Data Responsibly.

Some repositories have secure data services, to provide restricted, controlled access to data which may be sensitive or contain personally identifiable information. This can enable limited sharing and use by researchers.

If for reasons of sensitivity and/or proprietary agreements the full data cannot be shared, funders and journal publishers often make exceptions to their sharing policies. In such cases, it is still important to share as many related materials as possible so that others have clear instructions on how the data was obtained and analyzed (e.g., share code used for analysis and formally cite the detailed source of data in your publication).  

Generally, for data underlying a journal article, data should be published at the time of article publication (and cited within or linked from the article). Check your journal's author guidelines for further requirements (e.g., for specific data types).  

• Ensure that you deposit your data in a trustworthy repository, which makes your data widely discoverable and citable (e.g., via a DOI). 
• Cite your published data in your own journal publications, using the recommended citation from your data repository whenever possible.  
• Register for an ORCID to ensure all your research outputs can be centrally tracked. 

Each journal will have a unique data policy. Check its author guidelines for requirements covering

• Data citation
• Data sharing and availability (including specific data types, e.g., protein structures or DNA sequences)
• Metadata requirements like controlled keywords, ontologies (e.g., MeSH), persistent identifiers (e.g., DOI, ORCID, or RRIDs), or funding agency terminology (e.g., Open Funder Registry)

If the data used are available from a repository (whether you published or found the data there), the data repsository should provide a recommended citation. Make sure that you include additional identifiers for research outputs or resources (e.g., RRIDs) and consult the DataCite guide to data citation. 

If you are still unsure, contact the library at your institution as most of them provide citation assistance.