|
Abstract:
|
Orienting the causal relationship between a pair of traits is a fundamental task in scientific research with significant implications in practice, such as in prioritizing molecular targets and modifiable risk factors for developing therapeutic and interventional strategies for complex diseases. A recent method, called Steiger’s method, using a single SNP as an instrument variable (IV) in the framework of Mendelian randomization (MR), has since been widely applied. We consider the problem in the presence of confounding with GWAS summary data as for (two-sample) MR. While allowing for multiple and possibly correlated SNPs as IVs, we further relax the three IV assumptions in MR for valid causal inference. Through extensive simulations and real data application, we demonstrate the superior performance and robustness of our proposed method over Steiger’s method and bi-directional MR.
|
