To improve diagnostic yield for rare diseases, we developed the Human Phenotype Ontology (HPO) in 2008 as
a comprehensive bioinformatic resource that provides a standardized terminology of phenotypic abnormalities
for the analysis of human diseases. HPO reduces ambiguity in disease descriptions—thus enabling more robust
differential diagnosis and clinical care—and enables phenotypic contextualization of genomic data for
diagnostics and precision medicine.
The performance of computational algorithms for differential diagnostics with HPO terms depends critically on
the comprehensiveness and depth of HPO annotations for diseases. However, the current manual nature of our
biocuration process has limited the quality, depth, and coverage of these annotations. Therefore, this proposal's
objectives are to greatly expand the corpus of disease-phenotype annotations by automating portions of the
curation and expanding the computational disease model.
This project, HPO: Accelerating Computational Integration of Clinical Data for Genomics, will maintain and
advance HPO resources to address the needs of a growing number of medical disciplines that have adopted the
HPO. We will achieve this goal by 1) automating HPO development, maintenance, and release processes, 2)
developing representations of rare disease treatments and interventions, and 3) extending our current
computational disease models to represent time course, sex biases, and frequency of events, and to incorporate
case report data. We also provide a sustainable solution to community contribution with a user-friendly, web-
based portal to enable contributors to vet and suggest improvements to the ontology and the annotations and
grow the HPO contributor community.
In summary, our project addresses the most pressing needs for advancements of the HPO to ensure sustainable,
robust, and rigorous development, to enable HPO resources to support new communities, new applications, and
more medical disciplines.