The use of two-sample methods for Mendelian randomization analyses on single large datasets

Abstract

Abstract Background With genome-wide association data for many exposures and outcomes now available from large biobanks, one-sample Mendelian randomization (MR) is increasingly used to investigate causal relationships. Many robust MR methods are available to address pleiotropy, but these assume independence between the gene-exposure and gene-outcome association estimates. Unlike in two-sample MR, in one-sample MR the two estimates are obtained from the same individuals, and the assumption of independence does not hold in the presence of confounding. Methods With simulations mimicking a typical study in UK Biobank we assessed the performance, in terms of bias and precision of the MR estimate, of the fixed-effect and (multiplicative) random-effects meta-analysis method, weighted median estimator, weighted mode estimator and MR-Egger regression, used in both one-sample and two-sample data. We considered scenarios differing for: presence/absence of a true causal effect; amount of confounding; presence and type of pleiotropy (none, balanced or directional). Results Even in the presence of substantial correlation due to confounding, all methods performed well when used in one-sample MR except for MR-Egger, which resulted in bias reflecting direction and magnitude of the confounding. Such bias was much reduced in the presence of very high variability in instrumental strength across variants (I 2 GX of 97%). Conclusions Two-sample MR methods can be safely used for one-sample MR performed within large biobanks, expect for MR-Egger. MR-Egger is not recommended for one-sample MR unless the correlation between the gene-exposure and gene-outcome estimates due to confounding can be kept low, or the variability in instrumental strength is very high. Key Messages <jats:list list-type="bullet"><jats:list-item> Current availability of phenotypic and genetic data from large biobanks, such as UK Biobank, has led to increasing use of one-sample Mendelian randomization (MR) to investigate causal relationships in epidemiological research <jats:list-item> Robust MR methods have been developed to address pleiotropy, but they assume independence between the gene-exposure and gene-outcome association estimates; this holds in two-sample MR but not in one-sample MR <jats:list-item> We illustrate the practical implications, in terms of bias and precision of the MR causal effect estimate, of using robust two-sample methods in one-sample MR studies performed within large biobanks <jats:list-item> Two-sample MR methods can be safely used for one-sample MR performed within large biobanks, expect for MR-Egger regression <jats:list-item> MR-Egger is not recommended for one-sample MR unless the correlation between the gene-exposure and gene-outcome estimates due to confounding can be kept low, or the variability in instrumental strength is very high