Method | Assumptions | Strengths | Weaknesses | PubMed ID |
IVW | All genetic instruments are valid. | Has optimal statistical power. | Estimates are biased if there is directional pleiotropy. | 24 114 802 |
Weighted median | More than 50% of the weight comes from valid genetic instruments. | Informs the estimate supported by the majority of evidence. | May be less efficient. | 27 061 298 |
MR-PRESSO | The largest group of candidate instruments with similar estimates is the group of valid instruments. | Detects outliers and provides an estimate after removal of outliers | May have high false-positive rate. | 29 686 387 |
MR-Egger | Associations of genetic instruments with the exposure are uncorrelated with any pleiotropic effects of the instruments on the outcome (InSIDE assumption). | Quantifies directional pleiotropy and provides a consistent estimate even if all genetic variants are pleiotropic. | Has low precision and can be strongly influenced by outliers. | 29 040 600 |
IVW, inverse variance weighted; MR, Mendelian randomization; PRESSO, pleiotropy residual sum and outlier.