Leveraging terminological resources for mapping between rare disease information sources.

Area of Science:

• Biomedical Informatics
• Medical Terminology Management
• Rare Disease Data Integration

Background:

• Information sources for rare diseases lack consistent cross-referencing, hindering user navigation.
• Manual creation of cross-references is time-consuming and expensive.

Purpose of the Study:

• To automate the mapping between the Genetic and Rare Diseases (GARD) and Orphanet rare disease information sources.
• To leverage terminological resources, specifically the Unified Medical Language System (UMLS), for cross-referencing.

Main Methods:

• Rare disease terms from Orphanet and the Office of Rare Disease Research (ORDR) were mapped to the UMLS.
• The UMLS served as a central pivot to link the different rare disease terminologies.
• Results were validated against manually curated cross-references to the Online Mendelian Inheritance in Man (OMIM) database.

Main Results:

• The automated mapping achieved a precision of 94%, a recall of 63%, and an F1-score of 76%.
• The developed mapping facilitates more complete and consistent cross-references between GARD and Orphanet.
• The methodology is adaptable for other rare disease information sources.

Conclusions:

• Automated mapping using UMLS offers an efficient solution for interlinking rare disease databases.
• This approach enhances data accessibility and consistency for researchers and patients.
• The findings support the development of a more integrated rare disease information ecosystem.