MR-Clust: clustering of genetic variants in Mendelian randomization with similar causal estimates

Objectives:

The research aimed to utilize Mendelian randomization (MR) as a robust epidemiological technique to investigate and estimate causal relationships between specific risk factors and their resultant outcomes. This study sought to leverage genetic variants as instrumental variables to provide a reliable framework for understanding the underlying causal mechanisms of various health conditions. Ultimately, the goal was to contribute to scientific understanding of the factors contributing to disease causation and inform potential interventions or preventive measures.

Solutions:

• Incorporation of innovative features to accommodate different levels of uncertainty associated with each causal estimate.
• Identification of ‘null’ and ‘junk’ clusters to ensure recognition of genuine biological mechanisms and safeguard against random noise.
• Consistent performance in accurately identifying the correct number of variant clusters, surpassing other methods in simulation analyses.
• Effective application in real-world scenarios, providing insights into the impact of risk factors on outcomes and guiding scientific inquiry towards novel hypotheses.

Challenges:

• Divergence in causal estimates from genetic variants suggesting multiple distinct causal mechanisms.
• Identification of genuine biological mechanisms amid variability in causal estimates.
• Ensuring accuracy in cluster recognition while accounting for uncertainty.

Impact:

• Superior performance of MR-Clust over existing methods in identifying variant clusters.
• Consistent identification of correct number of variant clusters in simulation analyses.
• Demonstration of effectiveness in real-world scenarios, such as assessing the impact of blood pressure on coronary artery disease risk.
• Revelation of intriguing associations and generation of hypotheses, leading to deeper understanding of biological mechanisms.
• Guidance of scientific inquiry towards novel hypotheses and insights into complex interplay between risk factors and outcomes.