Phenotype validation in electronic health records based genetic association studies

Area of Science:

• Genomics
• Biostatistics
• Health Informatics

Background:

• Electronic health records (EHRs) linked with genotype data enable studying genetic susceptibility across diverse disease phenotypes.
• EHR-derived phenotype data, despite potential misclassification, are valuable for gene discovery, especially in phenome-wide association studies (PheWAS).
• Validating EHR-based findings with gold-standard chart reviews is crucial for reliable genetic association studies.

Purpose of the Study:

• To propose a genotype-stratified case-control sampling strategy for efficient phenotype validation.
• To develop statistical methods for analyzing combined validated and error-prone EHR data.
• To assess the power improvements and bias correction offered by the proposed strategy.

Main Methods:

• A novel genotype-stratified case-control sampling strategy for subject selection in phenotype validation.
• Development of a closed-form maximum-likelihood estimator for odds ratio parameters.
• Formulation of a score statistic for genetic association testing using mixed-quality phenotype data.

Main Results:

• The proposed genotype-stratified strategy maintains nominal type I error rates.
• This method significantly increases power for detecting genetic associations compared to EHR-only sampling.
• The strategy corrects bias in odds ratio estimates and reduces variance, particularly for low minor allele frequencies.

Conclusions:

• Genotype-stratified sampling enhances the reliability and power of genetic association studies using EHR data.
• The developed statistical methods effectively integrate validated and error-prone phenotype data.
• This approach offers a robust framework for discovering genetic susceptibility in large-scale biobanks.