Mendelian randomization with proxy exposures: challenges and opportunities.

Area of Science:

• Human genetics
• Complex trait analysis
• Biomarker discovery

Background:

• Identifying modifiable causal risk factors for complex diseases is crucial in human genetics.
• Mendelian randomization (MR) is a powerful tool for this, but horizontal pleiotropy presents a significant challenge.
• Previous MR studies on biomarkers like cholesterol and C-reactive protein have shown promise but also revealed complexities.

Purpose of the Study:

• To address the challenge of horizontal pleiotropy in Mendelian randomization.
• To illustrate how cis-MR with proxy exposures can overcome pleiotropy.
• To rediscover causal relationships in vitamin D synthesis and glycolysis using metabolite levels.

Main Methods:

• Utilized UK Biobank data for two case studies: glycolysis and vitamin D synthesis.
• Employed a cis-MR framework with metabolite levels (pyruvate, histidine) as proxy exposures.
• Analyzed variant effects on downstream metabolites to infer protein function perturbations.

Main Results:

• Demonstrated that measured metabolites (pyruvate, histidine) did not directly cause outcomes (red blood cell count, vitamin D level).
• Successfully used variant effects on metabolite levels as proxy exposures in cis-MR.
• Rediscovered the causal roles of histidine ammonia lyase (HAL) in vitamin D synthesis and glycolysis in red blood cell survival.

Conclusions:

• Cis-MR with proxy exposures can overcome horizontal pleiotropy in Mendelian randomization.
• This approach allows for the inference of causal relationships even when direct metabolite effects are absent.
• Highlights the assumptions and practical challenges for valid cis-MR inferences with proxy exposures.