Predicting disease-related phenotypes using an integrated phenotype similarity measurement based on HPO

Area of Science:

• Medical Informatics
• Bioinformatics
• Computational Biology

Background:

• Disease diagnosis efficiency relies on phenotype ontology analysis.
• Human Phenotype Ontology (HPO) is widely used for gene and disease identification via semantic similarity.
• Existing HPO methods overlook phenotype term definitions, limiting accuracy.

Purpose of the Study:

• Introduce DisPheno, a novel phenotype similarity measurement.
• Incorporate HPO term definitions, structure, and annotations for enhanced similarity.
• Improve disease identification efficiency.

Main Methods:

• Developed DisPheno algorithm for phenotype term similarity.
• Integrated phenotype term definitions into similarity calculations.
• Utilized gene and disease annotations for phenotype-set similarity.

Main Results:

• DisPheno effectively measures HPO-based phenotype semantic similarity.
• Demonstrated superior performance compared to five state-of-the-art methods.
• Showcased improved disease identification, particularly on noisy datasets.

Conclusions:

• DisPheno offers a more comprehensive approach to phenotype similarity.
• The method enhances the efficiency and accuracy of disease diagnosis.
• DisPheno is particularly beneficial for analyzing complex or noisy patient data.