The study addresses the issue of causal inference in the presence of measurement error, particularly in the context of mediation-based approaches and Mendelian randomization (MR). The authors show that measurement error can lead to incorrect causal direction in mediation-based methods like the Causal Inference Test (CIT), even with large sample sizes. They introduce an extension to MR, called MR Steiger, which can infer the causal direction between traits using summary data from genome-wide association studies (GWAS). This method is less susceptible to measurement error and unmeasured confounding. The authors apply MR Steiger to infer the causal direction between DNA methylation and gene expression levels, finding that DNA methylation is often the causal factor but that this conclusion is highly sensitive to measurement error. The study emphasizes the importance of combining multiple causal inference methods and conducting sensitivity analyses to robustly determine causal relationships.The study addresses the issue of causal inference in the presence of measurement error, particularly in the context of mediation-based approaches and Mendelian randomization (MR). The authors show that measurement error can lead to incorrect causal direction in mediation-based methods like the Causal Inference Test (CIT), even with large sample sizes. They introduce an extension to MR, called MR Steiger, which can infer the causal direction between traits using summary data from genome-wide association studies (GWAS). This method is less susceptible to measurement error and unmeasured confounding. The authors apply MR Steiger to infer the causal direction between DNA methylation and gene expression levels, finding that DNA methylation is often the causal factor but that this conclusion is highly sensitive to measurement error. The study emphasizes the importance of combining multiple causal inference methods and conducting sensitivity analyses to robustly determine causal relationships.