PheMap: a multi-resource knowledge base for high-throughput phenotyping within electronic health records

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Extracting phenotypes from electronic health records (EHRs) is complex and time-consuming.
• Automated phenotyping methods are crucial for large-scale genetic research.

Purpose of the Study:

• To develop and evaluate PheMap, a high-throughput phenotyping approach for EHRs.
• To streamline the extraction of research-quality phenotype data.

Main Methods:

• PheMap utilizes a knowledge base of medical concepts linked to phenotypes, derived from public resources.
• It searches EHRs for these concepts to calculate phenotype probabilities.
• PheMap was validated against clinician-validated algorithms from the eMERGE network for type 2 diabetes mellitus (T2DM), dementia, and hypothyroidism.

Main Results:

• PheMap achieved >97% accuracy for T2DM, dementia, and hypothyroidism phenotypes.
• PheMap-derived data replicated 43/51 previously reported disease-associated variants in GWAS.
• PheMap showed comparable or better performance than traditional methods in GWAS and PheWAS.

Conclusions:

• PheMap significantly simplifies phenotype extraction from EHRs.
• It offers comparable or superior performance to existing phenotyping strategies.
• PheMap is publicly accessible, facilitating broader research applications.