Improving Imputation Accuracy by Inferring Causal Variants in Genetic Studies

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genotype imputation is crucial for genome-wide association studies (GWAS), enhancing power and aiding interpretation of untyped variants.
• Existing imputation methods often assume no causal variants within a locus, which can lead to overly conservative predictions of association statistics.

Purpose of the Study:

• To develop and evaluate CAUSAL-Imp, a novel imputation method that accounts for potentially causal variants within a locus.
• To address the underestimation of association statistics by traditional imputation methods when causal variants influence the trait.

Main Methods:

• CAUSAL-Imp builds upon existing GWAS imputation techniques that leverage the multivariate normal distribution of marginal statistics.
• The method incorporates information about variants within a locus that may directly affect the studied trait.
• Performance was assessed using both simulated and real-world genetic datasets.

Main Results:

• Traditional imputation methods provide overly conservative estimates of association statistics for trait-influencing variants.
• CAUSAL-Imp demonstrates reduced bias in estimating association statistics at untyped variants, particularly when causal variants are present.
• The proposed method improves the accuracy of imputation for variants associated with the trait of interest.

Conclusions:

• CAUSAL-Imp offers a more accurate approach to genotype imputation in GWAS by considering locus-specific causal variants.
• This method enhances the reliability of association statistics for untyped variants, improving GWAS interpretation.
• The findings suggest that accounting for potential causality within loci is essential for precise genetic association analyses.