Search for a command to run...
This FAIRsharing record describes: Term mappings are of fundamental importance to interoperability, yet often lack metadata to be correctly interpreted and applied in contexts such as data integration or transformation. For example, are two terms equivalent or merely associated? Are they narrow or broad matches? etc. Such relationships between the mapped terms often remain unclear, which makes them very hard to use in scenarios that require a high degree of precision (such as diagnostics or risk prediction). Furthermore, the lack of metadata on the methods and rules involved in producing the mappings and confidence estimations regarding their correctness makes it hard to combine and reconcile mappings, especially curated and automated ones. Working as part of a collaborative group, we have developed a Simple Standard for Sharing Ontological Mappings (SSSOM) which addresses these problems by introducing a simple vocabulary for mapping metadata and defining an easy to use table-based format that can be integrated into regular data science pipelines without the need to parse or query ontologies defining a set of exports formats such as RDF/XML and JSON-LD. SSSOM is defined using a LinkML schema (https://linkml.github.io), and defines metadata for many key features of term mappings and mapping sets, such as mapping confidence, versioning, mapping tools and match types (lexical, logical, human-curated). The SSSOM specification can be found at http://w3id.org/sssom. Associated toolkits are being developed, for example https://github.com/mapping-commons/sssom-py and https://incenp.org/dvlpt/sssom-java/.