Analysis in case-control sequencing association studies with different sequencing depths

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Next-generation sequencing (NGS) provides high-quality data but is expensive for large-scale studies.
• Combining study-specific case data with public control data is a cost-saving strategy.
• Systematic differences in sequencing quality, like depth, can exist between case and control datasets.

Purpose of the Study:

• To develop statistical methods for genetic association studies using combined datasets with differential sequencing quality.
• To account for systematic differences in sequencing errors between cases and controls.
• To enable robust analysis while adjusting for covariates like population stratification.

Main Methods:

• Proposed a regression calibration (RC)-based method.
• Developed a maximum-likelihood method.
• Both methods account for differential sequencing errors and adjust for covariates.

Main Results:

• Both proposed methods control type I error rates effectively.
• The methods demonstrate comparable power to analyses using true genotypes, even with varying sequencing depths.
• The RC method can utilize naive variance estimates and standard software in certain scenarios.

Conclusions:

• The developed RC and maximum-likelihood methods are effective for association studies with combined case-control data of differing quality.
• These methods provide a viable solution for cost-effective genomic studies.
• The approach was validated through simulations and application to acute lung injury exome data.