Skip to content

sepio-linkml

SEPIO was first developed as an ontology by the Monarch Initiative to support standardized RDF representations of evidence and provenance across integrated genotype-phenotype datasets (the Scientific Evidence and Provenance Information Ontology). The ontological model has since been abstracted into a generic Core Information Model (IM) that can be implemented in any language or format. The Core IM is domain-agnostic, and able to represent assertions and their evidence and provenance of any kind. Application of SEPIO to a specific data set or use case requires defining a ‘Profile’ that extends/customizes the generic core model for a particular domain or application.

The components of the SEPIO Framework include:

  1. A Domain Analysis Model (DAM): an informal description of the domain we are modeling (scientific assertions and their evidence/provenance)
  2. A Core Information Model (IM): defines data structures that can represent information about this domain (for any type of assertion and evidence).
  3. A 'Profiling' Methodology: Implementations extend the core model with domain-specific content to define a “SEPIO Profile” - a custom schema for a particular application or use case.
  4. Ontology Support: An ontological representation of the core model that can be used if desired to produce linked data with ontology-based semantics.

The modeling framework-based approach addresses challenges posed by the diversity of types, levels of complexity, and use cases for evidence and provenance across knowledge domains and application - which means there is no ‘one-size-fits-all’ solution. A framework allows custom models built on a common semantic foundation can provide a base level of understanding and interoperability, without restricting expressivity. While this approach may not always support out-of-the-box interoperability across all communities of use, it can significantly lower barriers to aggregating, harmonizing, and operating across disparate data.

The SEPIO Core-IM specified in this repository was written in a yaml format using the LinkML modeling language, and uses LinkML tooling to automatically generate formal json schema specification and Github pages web documentation found here.