An Introduction to the Human Phenotype Ontology

The human phenotype ontology (HPO) intends to offer a tool that will allow large-scale computational analysis of the human phenome. The HPO currently contains over 9500 terms, each of which describes an individual phenotypic anomaly. The terms are arranged in a directed acyclic graph and are connected by “is-a” edges, such that a term represents a more specific or limited instance of its parent term(s). All relationships in the HPO are is_a relationships, i.e. a simple class-subclass relationships. For instance, Abnormality of the feet is_a Abnormality of the lower limbs. The relationships are transitive, meaning that they are inherited up all paths to the root. Organ abnormality is the main subontology of the HPO and contains descriptions of clinical abnormalities. Additional subontologies are provided to describe inheritance patterns and onset/clinical course.

What is An Ontology and What Use is An Ontology for Medical Genetics Research?

The word ontology is derived from Greek words meaning the study of existance and being. More recently, the word ontology has been used in computer science to describe systems that describe concepts within some domain and relationships between those concepts. The Gene Ontology consortium has developed an extensive ontology describing molecular functions, biological processes, and cellular locations over the last decade and a number of groups have supplied annotations using the GO terms to gene products of many organisms.

The study of human phenotypes in the context of hereditary disease has the potential to lead to great insight on the function of genes and genetic networks. The human phenotype ontology (HPO) intends to offer a computational tool that will allow large-scale computational analysis of the human phenome.

What is the Medical Focus of the Human Phenotype Ontology?
Currently, the most useful and comprehensive database of human hereditary disorders is the Online Mendelian Inheritance in Man (OMIM) database. OMIM contains clinical data on several thousand primarily monogenic hereditary disorders. Since OMIM and its predecesor Mendelian Inheritance in Man (McKusick) were initially developed over 30 years ago, OMIM does not use a controlled vocabulary to describe clinical features, and uses a more or less simple hierarchical scheme to assign clinical features to organ systems. Despite these drawbacks, OMIM was and will continue to be an incredibly useful resource for clinicians and researchers. We have therefore mapped nearly all clinical descriptions in OMIM that are used more than once to terms of the HPO. We have embedded these mappings into an ontological, multitiered structure that is described below and intend to continue refining and improving the HPO. In the future, the annotations to the HPO will be extended to include chromosomal disorders.

 

Terms in the Human Phenotype Ontology
Each term in the HPO describes a clinical abnormality. These may be general terms, such as Abnormality of the musculoskeletal system or very specific such as Chorioretinal atrophy. Each term is also assigned to one of the three ontologies, Organ abnormality, Inheritance and Onset and Clinical course. Most of the terms of the HPO belong to the Organ abnormality ontology. The terms have an ID such as HP:0001140 and a name such as "Epibulbar dermoids".  Some terms have definitions such as "An epibulbar dermoid is a benign tumor typically found at the junction of the cornea and sclera (limbal epibullar dermoid). " The source of the definition must be indicated. For the moment, the source is indicated as HPO:curators. Many terms have synonyms. Most of the synonyms are taken from the OMIM entries, which were made over several decades without a controlled vocabuly. For instance, "Epibulbar dermoid" is taken to be a synonym of "Epibulbar dermoids". The synonyms may also be useful for searches.

 

The Organ abnormality ontology
This is the main ontology of the HPO and contains descriptions of clinical abnormalities. The level 1 children of Organ abnormality is formed by terms such as Abnormality of the musculoskeletal system and Hematological abnormality.

 

The Inheritance ontology
This relatively small ontology is intended to describe the mode of inheritance and contains terms sich as Autosomal dominant.

 

The Onset and Clinical course Ontology
This ontology describes the typical time of onset of clinical symptoms and their speed of progression. It contains terms such as Onset in childhood. and Rapidly progressive.

 

The Structure of the Human Phenotype Ontology
The ontologies are structured as directed acyclic graphs (DAG), which are similar to hierarchies but differ in that a more specialized term (child) can be related to more than one less specialized term (parent).  Cycles (cyclic paths in the graph) are not allowed. The relationship of the terms of the HPO to one another is displayed in the DAG. For instance, the term Aplasia/Hypoplasia of metatarsal bones is a child of both  Aplasia/Hypoplasia involving bones of the feet and Abnormalities of the metatarsal bones. The ability to encode multiple parents in a DAG adds to the flexibility and descriptiveness of the ontology. For instance, it is possible to search for all terms involving Aplasia/Hypoplasia of the skeleton as well as to search for all terms involving abnormalities of the foot. This would not be possible with a simple hierarchical system.

At the moment, all relationships in the HPO are is_a relationships,  i.e.  asimple class-subclass relationships. For instance, Abnormality of the feet is_a Abnormality of the lower limbs. The relationships are transitive, meaning that they are inherited up all paths to the root. For instance,
Abnormality of the lower limbs is_a Abnormality of the extremities, and thus Abnormality of the feet also is Abnormality of the extremities.