The Human Phenotype Ontology (HPO) provides a standardized vocabulary for describing the phenotypic abnormalities encountered in human disease. With over 18,000 terms covering clinical features from broad categories to highly specific manifestations, HPO has become an essential resource in rare disease diagnostics, genomic medicine, and clinical research.
Background
HPO was created in 2008 at the Charite - Universitatsmedizin Berlin by Peter Robinson and colleagues, motivated by the need for a computationally tractable vocabulary to describe clinical phenotypes. Before HPO, phenotypic descriptions in the medical literature and clinical databases were largely unstructured, making it difficult to systematically compare phenotypes across patients or to link clinical features to underlying genetic causes.
The ontology was developed using data from Online Mendelian Inheritance in Man (OMIM) and the medical literature. It is now maintained as part of the Monarch Initiative, a collaborative effort to integrate biological information for translational research, with primary development at the Jackson Laboratory for Genomic Medicine.
Purpose & Scope
HPO provides a formal representation of phenotypic abnormalities observed in human disease. Its scope includes:
- Phenotypic features organized hierarchically from general (e.g., "Abnormality of the nervous system") to specific (e.g., "Cerebellar ataxia")
- Annotations linking HPO terms to hereditary diseases (over 156,000 annotations)
- Frequency information indicating how often a phenotype is observed in a given disease
- Cross-references to other ontologies and medical terminologies
The ontology is structured as a directed acyclic graph, with the primary axis organized by anatomical system. A separate sub-ontology covers modes of inheritance.
Key Statistics
| Metric | Value |
|---|---|
| HPO terms | 18,000+ |
| Disease annotations | 156,000+ |
| Gene-phenotype links | Extensive |
| Sub-ontologies | Phenotypic abnormality, Mode of inheritance, Clinical modifier |
Serializations & Technical Formats
HPO is distributed in OBO format, OWL (Web Ontology Language), and JSON. The canonical namespace URI is http://purl.obolibrary.org/obo/HP_, and individual terms follow the pattern HP:0000001. Release files are available from the HPO website and through the OBO Foundry.
Governance & Maintenance
HPO is maintained by the Monarch Initiative team at the Jackson Laboratory for Genomic Medicine, with contributions from a global community of clinicians and geneticists. The ontology undergoes regular releases with new terms, revised definitions, and updated disease annotations. HPO is a member of the OBO Foundry, which provides standards for ontology development and interoperability.
Notable Implementations
HPO is a core component of clinical genomics platforms including Exomiser (variant prioritization tool), Phenomizer (differential diagnosis tool), and the Monarch Initiative knowledge graph. It is used by ClinVar, DECIPHER, Orphanet, and many clinical genetics laboratories for structured phenotype recording. The GA4GH Phenopackets standard uses HPO terms as its primary phenotype vocabulary.
Related Standards
- Gene Ontology — complementary ontology describing gene function rather than clinical phenotype
- Disease Ontology — ontology of human diseases, closely integrated with HPO annotations
Monarch Initiative