search

cran / MendelianRandomization

:exclamation: This is a read-only mirror of the CRAN R package repository. MendelianRandomization — Mendelian Randomization Package
42 stars 26 forks source link

absolute value of Bx #3

Open ghost opened 3 years ago

ghost commented 3 years ago

I'm calculating Isq using mr_egger(). But I noticed that you took absolute value of Bx in line 70. Could you explain why did you take the absolute value? There is no absolute value in the formula of Q statistic proposed by Bowden IJE 2016. Thank you.

gaborcsardi commented 3 years ago

Hi, this is a read only mirror of CRAN, please see the package authors in the DESCRIPTION file. Thanks!

sb452 commented 1 year ago

MR-Egger depends on the orientation of the genetic variants - i.e. which is the effect allele, which is the other (non-effect) allele. Conventionally, we set all the effect alleles to be the exposure-increasing alleles.