A0849
Title: Simultaneous inference of causal existence and direction via the coefficient of determination
Authors: Huiling Liao - Illinois Institute of Technology (United States) [presenting]
Abstract: In Mendelian randomization, two single SNP-trait correlation-based methods have been developed to infer the causal direction between an exposure (e.g., a gene) and an outcome (e.g., a trait), called MR Steiger's method and its recent extension called causal direction-ratio (CD-Ratio). Motivated by the limited variance explained by individual SNPs and potential linkage disequilibrium among SNPs, a new method is introduced that leverages R2, the coefficient of determination, to aggregate information from multiple, possibly correlated, SNPs. The approach extends the existing method to accommodate multiple instrumental variables (IVs), enabling joint inference of both the existence and direction of causality simultaneously. This is particularly advantageous in transcriptome-wide association studies (TWASs), where expression datasets often have limited sample sizes. The method can be applied to GWAS (genome-wide association study) summary statistics using a suitable reference panel. To address the issue of invalid IVs, a complementary procedure is also proposed, R2S, which identifies and removes such IVs to improve robustness and other potential extensions.
