Exploring and Accounting for Genetically Driven Effect Heterogeneity in Mendelian Randomization.

Area of Science:

• Epidemiology
• Statistical Genetics
• Pharmacogenetics

Background:

• Mendelian randomization (MR) estimates causal effects using genetic variants as instrumental variables.
• A key MR assumption is homogeneity, meaning causal effects don't vary across instrument levels.
• Pharmacogenetics actively seeks genetically driven effect heterogeneity for precision medicine.

Purpose of the Study:

• To combine the Triangulation WIthin a STudy (TWIST) framework with MR for analyzing average causal effects and genetic heterogeneity.
• To propose two new methods for estimating genetically driven effect heterogeneity.
• To estimate causal effects separately in genetic groups with and without risk alleles.

Main Methods:

• Utilized a combination of traditional MR and the TWIST framework.
• Developed two novel methods incorporating homogeneity-respecting and homogeneity-violating genetic variants.
• Applied methods to ALSPAC study data examining smoking in pregnancy and offspring birth weight.

Main Results:

• The study successfully estimated genetically driven effect heterogeneity in the causal effect of smoking on birth weight.
• Causal effects were estimated separately for mothers with different genetic predispositions to quitting smoking.
• Demonstrated the utility of the new methods in a real-world observational dataset.

Conclusions:

• The developed methods effectively characterize both average causal effects and genetically driven effect heterogeneity.
• This integrated approach advances precision medicine by identifying subgroups with differential treatment effects.
• The findings provide insights into the complex interplay between genetics, smoking, and offspring birth weight.