The schemas used by BELBio for important data objects are managed here. The schemas are in JSONSchema format, but the original and editable versions are stored in YAML format to allow for additional comments and easier reading of the data objects.
The data object schemas are for:
- nanopubs
- edges
- terminologies
- orthologies
- taxonomy
Tooling is also available including a tool to convert the YAML formats to JSON.
Many of the following schemas will support JSONLines (jsonlines.org) format to allow streaming of the dataset so one doesn't have to load all of the JSON file before processing. One can load each sub-object, e.g. a term at at time in the terminology schema which for some terminologies with a few million entries can make an enormous different in memory usage.
The nanopub schema is for capturing curated knowledge or assertions as a triple (subject, relation, object or SRO). The triple (or array of triples) have a citation indicating from where the triple was sourced, context such as for BEL Nanopubs what tissue, disease context, cell line, species, etc the triples are associated with. Miscellaneous metadata is also captured for the triple.
The objects imported/exported to/from the EdgeStore are in this format defined by the edge schema.
Terms for the BEL Namespaces are loaded from datafiles in this format
Orthology data for BEL is loaded from datafiles in this format
Taxonomy data is loaded into BEL from datafiles in this format.
![@dependabot[bot]](https://avatars.githubusercontent.com/in/29110?s=64&v=4){kind=small}