Statistical inference in two-sample summary-data Mendelian randomization using robust adjusted profile score

Summary

This methodological paper advances statistical inference for two-sample summary-data Mendelian randomisation by developing estimators that account for pleiotropy—where genetic variants affect outcomes through multiple pathways, violating causal assumptions. The authors propose a robust adjusted profile score method that maintains consistency and asymptotic normality when both systematic and idiosyncratic pleiotropy are present, validated through simulation and real epidemiological datasets. The work addresses a critical challenge in using genetic variation for causal inference in population science.

UK applicability

The statistical methods developed are applicable to UK-based genetic epidemiological studies and biobanks (such as UK Biobank) that employ two-sample Mendelian randomisation designs. The robustness improvements may enhance the reliability of causal inference conclusions drawn from UK health and agricultural genomic datasets.

Key measures

Maximum profile likelihood estimators, adjusted profile scores, asymptotic normality, consistency of causal effect estimates under pleiotropy

Outcomes reported

The study developed and evaluated statistical methods for two-sample summary-data Mendelian randomisation, addressing pleiotropy bias in causal effect estimation. The methods were assessed using simulated and real genetic datasets to demonstrate robustness and efficiency.

Topic tags