absolute value of Bx

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]

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

[sb452]

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.