Summary
This methodological paper proposes the mode-based estimate (MBE), a novel Mendelian randomisation method for summary data that relaxes instrumental variable assumptions by requiring only that the largest cluster of similar causal effect estimates derives from valid instruments, even if the majority are invalid. Through simulations and application to lipid and urate phenotypes, the authors demonstrate that MBE exhibits lower bias and type-I error than competing methods under the null hypothesis, though with reduced statistical power. The method is proposed as a sensitivity analysis tool to be used alongside established approaches such as inverse variance weighting and weighted median estimation.
UK applicability
This is a methodological contribution to epidemiological causal inference techniques applicable to UK researchers conducting Mendelian randomisation studies using publicly available GWAS summary data. The method's relaxed assumptions may improve robustness of causal estimates in UK biobank and consortia analyses, though power limitations noted in the paper warrant consideration in study design.
Key measures
Bias, type-I error rates, statistical power, and sample size requirements of the mode-based estimate compared with inverse variance weighting, weighted median, and MR-Egger regression methods
Outcomes reported
The study evaluated the performance of a novel mode-based estimate (MBE) method for obtaining causal effect estimates from multiple genetic instruments in Mendelian randomisation analyses. It demonstrated the method's application to investigating causal effects of plasma lipid fractions and urate levels on coronary heart disease risk.
Topic tags
Dig deeper with Pulse AI.
Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.