Gene Regulation Ontology
Ontological resource from the BOOTStrep project for the representation of gene regulation events
The Gene Regulation Ontology (GRO) is a conceptual model for the
domain of gene regulation. It covers processes that are linked to the
regulation of gene expression as well as physical entities that are
involved in these processes (such as genes and transcription factors) in
terms of ontology classes and semantic relations between classes. GRO
is intended to represent common knowledge about gene regulation in a
formal way rather than representing extremely fine-grained classes as
can be found in ontologies such as the Gene Ontology (GO) (created for
data base annotation purposes) and various relevant databases. The
main purpose of the ontology is to support NLP applications. It has a
particular focus on the relations between processes and the molecules
(participants) involved. The basic structure of the GRO is a direct
acyclic graph (DAG) with ontology classes as nodes and is-a relations
between classes as edges. The taxonomic backbone is further enriched
by several semantic relation types (part-of, from-species,
participates-in with the two sub-relations agent-of and patient-of).
Poster at ISMB Bio-Ontologies SIG Workshop 2007: Bio-Ontologies in the context of the BOOTStrep project
Beisswanger,E., Lee,V., Kim,J.J., Rebholz-Schuhmann,D., Splendiani,A., Dameron,O., Schulz,S., Hahn,U. Gene Regulation Ontology (GRO): Design Principles and Use Cases. Stud Health Technol Inform. 2008;136:9-14.
GRO version 0.5 released. It contains 70 new classes and 6 new object properties. 3 classes from version 0.4 (GRO_v0.4
) were deleted. Class GRO:Affecting and subclasses are subject to change (GRO_v0.5
Almost all classes have now a textual description in the new version (GRO_v0.3
The downloaded OWL file needs to be classified by a reasoner to infer all class relations inferable from the source class definitions. In case you are unaware of reasoners look for pellet, racer, kaon2 or fact.
GRO is implemented in OWL DL. A typical GRO class has the follwing features:
ID (exactly one)
: a unique identifier starting with the namespace of
Name (exactly one): a name in natural language
Definition (exactly one): a textual definition
Parents (one to many): at least one parent class
Relations (zero to many): semantic relations linking the class to
other ontology classes
References (zero to many): references to similar terms in external
Synonyms (zero to many): synonyms of the term name
Use of pre-existing resources
The GRO has been created reusing terms from external ontological
resources. All terms taken from an external resouce provide a
reference pointing back to it. These are the external resources the GRO has been populated from:
Gene Ontology (GO)
, (molecular functions,
biological processes, cellular components, such as TF activity,
repressor activity, transcription, regulation of transcription,
Sequence Ontology (SO)
regions and attributes of sequence region, such as gene, TF binding site,
promoter, DNA, RNA)
, (eukaryotes: fungus, plant, protist,
prokaryotes: bacterium, archaebacterium)
, (chemical entities)
The Ontology serves as a communication means:
Other use of the ontology:
It conveys concepts involved in gene regulation
It conveys attributes and functions attached to parts of the gene regulatory process
It models gene regulation and describes the characteristics of a gene, protein, DNA and mRNA
Use the ontology as a representation means for gene regulatory information in a database.
The is-a relation (abstraction) can be used to improve concept identification in the text.
The part-of relations can improve concept identification in the text, i.e. if a part is described in the text the author might make reference to the whole concept (eventually improving disambiguation).
The agent-patient role pairs can be identified in the literature or text mining solutions can be constraint to findings of the proper agent-patient role.
The located-in role in terms of an agent-patient role can be used to identify mentions of such locations.
The ontological design was mainly steared by:
with contributions from: