Ontological background

Established biomedical ontologies define most terms for entity classes and predicates used in MolMeDB RDF. Other terms have been newly defined in MolMeDB Vocabulary as subclasses of terms from BioAssay Ontology or Units of Measurement Ontology. All ontologies and terms used in MolMeDB RDF dataset are named by IRI identifiers (listed below in Table 1). Each term has a unique identifier, but it is expected that terms from the same ontology share the same prefix. Large or more complex ontologies and datasets sometimes use collections of multiple prefixes for different domains.

Example

- Terms from MolMeDB Vocabulary use the prefix https://rdf.molmedb.upol.cz/vocabulary#.
- The full IRI for a membrane model is https://rdf.molmedb.upol.cz/vocabulary#MembraneModel, and for a penetration barrier it is https://rdf.molmedb.upol.cz/vocabulary#PenetrationBarrier.

Full IRI identifiers can be quite long and too cumbersome to use in texts and SPARQL queries, so it is common to use CURIE compact syntax, which allows for definition of namespace names as shorthands for prefixes. Compact IRI then consists of a namespace name in the place of a prefix, colon (:) separator and a local part. This documentation uses compact IRIs in most examples.

Example

For MolMeDB vocabulary namespace the name mmdbvoc is used in this text. Compact IRI for membrane model is mmdbvoc:MembraneModel, where mmdbvoc stands for the full prefix of the ontology (https://rdf.molmedb.upol.cz/vocabulary#) and MembraneModel is the local part identifying this particular term.

Note that compact IRIs alone are NOT valid IRIs. Any RDF document or SPARQL query using compact syntax needs to define namespace names and their corresponding prefixes. Example of mmdbvoc namespace name definition for MolMeDB Vocabulary in a SPARQL query :

...
PREFIX mmdbvoc: <https://rdf.molmedb.upol.cz/vocabulary#>
...

...
PREFIX mmdbvoc: <https://rdf.molmedb.upol.cz/vocabulary#>
...

Table 1: Ontologies, their namespaces and terms used in MolMeDB RDF.

MolMeDB Vocabulary

MolMeDB vocabulary has been constructed for needs of MolMeDB RDF where needed terms have not been found in already existing ontologies. Terms in the vocabulary can be divided into 4 categories - membrane model types, method types, assay endpoint (result) types and units. First three categories are defined as extensions of BioAssay Ontology (BAO) with relevant BAO terms as root classes and units as extension of Units of Measurement ontology (UO), with “unit” (obo:UO_0000000). This dictionary is not meant to be used in other projects and its terms are intended to be replaced later with relevant terms from BAO and UO (or different estabilished ontology) when they will be introduced.

Membranes and Methods

Both membrane model and method categories have more specific subclasses which are defined in MolMeDB vocabulary for greater ease of querying of assays using similar, but not same methods or membranes. Terms for these types are listed in Table 2 and Table 3. The hierarchy of these terms is designed to represent membrane and method typology used in MolMeDB website (membranes and methods). In the case of membranes we are developing a more robust Membrane Ontology (MEMON) to be used in future versions of MolMeDB dataset.

Table 2.  Membrane subtypes.

Table 3. Method subtypes.

Results (assay endpoints)

Results of assays are in BAO represented as endpoints and this ontology defines great number of them, however we needed to define most of endpoint terms for representation of MolMeDB dataset ourself. Endpoint terms defined in MolMeDB vocabulary and imported into it are in Table 4.

Table 4. Assay endpoint (result) classes.

Units

New units have been defined for use with some of assay endpoints. Hierarchy of defined units is listed in Table 5

Table 5. Units of measurement