Phenotypically Similar Rare Disease Identification from an Integrative Knowledge Graph for Data Harmonization:...

Area of Science:

• Genomics and Bioinformatics
• Medical Informatics
• Rare Disease Research

Background:

• Standardized protocols for rare disease data harmonization are lacking.
• Existing resources exhibit data redundancy and inconsistency.
• This hinders clinical decision-making and education in rare diseases.

Purpose of the Study:

• To systematically identify phenotypically similar Genetic and Rare Diseases (GARD).
• To support rare disease data harmonization through knowledge graph analysis.
• To determine similarity types among GARD diseases.

Main Methods:

• Programmatic identification of phenotypically similar GARD diseases.
• Comparison of disease mappings between GARD and other rare disease resources.
• Derivation of clinical manifestations from disease classifications and prioritization based on phenotypes and genotypes.

Main Results:

• Validated 87% of identified phenotypically similar disease pairs.
• Achieved 94% precision and 86% F-measure in similarity identification.
• Identified 662 phenotypically similar disease pairs for GARD data harmonization.

Conclusions:

• Successfully identified phenotypically similar rare diseases using two distinct approaches.
• Results will guide GARD data harmonization and expand translational science.
• Enhanced data transparency and consistency across rare disease resources.