A focusing framework for testing bi-directional causal effects in Mendelian randomization.

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Mendelian randomization (MR) is a key method for inferring causal effects using genetic variants as instrumental variables.
• Standard MR methods often struggle with bidirectional causal relationships and pleiotropic genetic variants (those affecting multiple traits).
• Existing MR assumptions can be overly restrictive or unachievable in complex scenarios with potential reverse causation.

Purpose of the Study:

• To develop a robust framework for testing bidirectional causal effects in Mendelian randomization.
• To address challenges posed by pleiotropic genetic variants in causal inference.
• To enhance the applicability and reliability of MR methods in complex biological systems.

Main Methods:

• Proposed a novel focusing framework designed to test bidirectional causal effects.
• Integrated the new framework with existing state-of-the-art Mendelian randomization techniques.
• Employed theoretical analysis to establish guarantees for the proposed method.

Main Results:

• Demonstrated that common MR assumptions are often untenable for bidirectional relationships.
• The proposed focusing framework effectively facilitates bidirectional causal effect testing.
• Validated the framework's performance through simulations and real-world data analyses.

Conclusions:

• The new focusing framework offers a more flexible and reliable approach to Mendelian randomization.
• This method advances causal inference by accommodating bidirectional relationships and pleiotropy.
• The findings have significant implications for genetic epidemiology and understanding complex trait causality.