Mendelian randomization analysis using mixture models for robust and efficient estimation of causal effects.

Egger regression and MR-mode yielded substantially wider confidence intervals thus leading to statistically non-significant or borderline significant results.

also in 11,409 other papers

The estimates from all the other methods were also in the negative direction, but those obtained from the weighted mode and Egger regression did not reach statistical significance due to large confidence intervals.

also in 111,027 other papers