MR Dictionary

Bayesian approach that allows for horizontal pleiotropy and reverse causality in scenarios where the direction of the causal effect between two traits is unknown.

BayesMR estimates the magnitude and likely direction of the causal effect between two traits by using a nested sampling scheme to compute the model evidence for both causal directions. By using a Bayesian model averaging approach, the model allows the determination of how likely the inferred causal direction is compared to that which is due to reverse causality.

References

Other terms in 'Pleiotropy-robust two-sample MR methods':