Integrating External Controls by Regression Calibration for Genome-Wide Association Study

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) identify disease-associated genetic variants.
• Large sample sizes increase statistical power in case-control studies but are costly.
• Integrating external control data is a cost-effective strategy to enhance power.

Purpose of the Study:

• To develop a robust method for integrating external control data in GWAS.
• To address the issue of inflated type I error rates due to batch effects.
• To improve statistical power in case-control association studies.

Main Methods:

• Propose an approach: integrating External Controls into the Association Test by Regression Calibration (iECAT-RC).
• iECAT-RC accounts for systematic differences (batch effects) between studies.
• Applied iECAT-RC to UK Biobank data for M72 Fibroblastic disorders, using genotype calling as the batch effect.

Main Results:

• Extensive simulations demonstrate iECAT-RC controls type I error rates.
• iECAT-RC boosts statistical power across all tested models.
• The method identified significant SNPs for fibroblastic disorders, showing higher detection probability in unbalanced studies.

Conclusions:

• iECAT-RC provides a reliable method for integrating external controls in GWAS.
• The approach effectively mitigates batch effects, enhancing study power and accuracy.
• iECAT-RC is particularly beneficial for unbalanced case-control association studies.