Every experiment and analysis we add to the portal needs to have some basic metadata included along with it. This metadata is shown when users view the experiment and used to help categorize and filter experiments so users can more easily find the kind of data they are interested in.
Experiment Metadata is meant to be independent of any specific analysis. This includes things like a name, description, biosample, assay, tested element type, and location of a file with the locations of the tested elements.
This metadata should be a json file called experiment.json and must adhere to
this schema.
Analysis Metadata is meant to cover information about a specific analysis. This includes things like a name, description, p-value threshold, p-value adjustment method, and the location of the file with the analysis results.
This metadata should be a json file called analysis001.json and must adhere to
this schema.
We have a set of data utilities that you can use to validate your data and metadata files. To validate metadata run the included metadata validation script that can be used for both experiment and analysis metadata
For more information on JSON Schemas, check out the specification. There are quite a few validators available. Please pick one that meets your needs. Our project uses jsonschema with no problems.