Table 1 MR methods assessed in this study.

From: Mendelian randomization while jointly modeling cis genetics identifies causal relationships between gene expression and lipids

MR method

Data required

Pleiotropy correction

LD correction

Minimum no. of IVs required

MR-link

Sumstats: E; ILD: O

Yes

Yes

1

Inverse variance weighted (IVW)28

Sumstats: O, E

No

No

1

LDA-MR-Egger17

Sumstats: O, E, LDR

Yes

Yes

3

MR-Egger18

Sumstats: O, E

Yes

No

3

MR-PRESSOa,19

Sumstats: O, E

Yes

No

3a

  1. MR methods assessed in our simulation with information about the type of data needed to make a causal estimate, the ability to correct for pleiotropy or linkage disequilibrium, and the minimum number of instrumental variables required.
  2. IVW inverse variance weighting, Sumstats summary statistics, ILD individual-level data, O outcome, E exposure, LDR linkage disequilibrium reference panel. For more details, see “Methods”.
  3. aWe refer here to the MR-PRESSO test that reports estimates after identifying and removing outliers, as the test without outliers generalizes to IVW estimates.
Back to article page