Title: Getting to GRIPs with MR-Egger 

Abstract: It is well known that the results of MR-Egger are sensitive to which alleles are designated as the effect alleles.  A pragmatic convention is to orient all SNPs to have positive effects on the exposure, which has some advantages in interpretation but also brings some philosophical limitations.  I will review these issues and propose a modification to MR-Egger which has the Genotype Recoding Invariance Property (GRIP).  This appears to be less biased than IVW and MR-Egger when there is directional pleiotropy under the “wrong” coding, and it is unbiased under a “Variance independent of covariance explained” condition (VICE). 

Biography: Frank Dudbridge is Professor of Statistical Genetics at the University of Leicester, and a Visiting Professor at the University of Bristol.   He has worked on statistical methods for genetic epidemiology, including family-based association studies, genome-wide association studies, Mendelian randomisation and polygenic risk prediction.  His applied collaborations span cardiovascular, respiratory, psychiatric and cancer genomics.

All welcome

In OS6 Oakfield House or Teams