Pulse Brain · Growing Health Evidence Index
Tier 3 — Observational / field trialPeer-reviewed

Statistical methods for <i>cis</i>‐Mendelian randomization with two‐sample summary‐level data

Apostolos Gkatzionis, Stephen Burgess, Paul J. Newcombe

Genetic Epidemiology · 2022

Read source ↗ All evidence

Summary

This methodological paper reviews and evaluates statistical approaches for conducting cis-Mendelian randomisation analyses with correlated genetic variants from a single gene region using summary-level data. Through simulation and applied case studies, the authors demonstrate that factor analysis and Bayesian variable selection produce more reliable causal inferences than simple pruning methods when weak instrument bias is suspected, although methods perform comparably under conditions of large sample sizes and strong genetic instruments. The work addresses a key technical challenge in genomic epidemiology: the numerical instability that arises when inverting ill-conditioned genetic correlation matrices in multi-variant analyses.

UK applicability

The methodological advances described may be relevant to UK genetic epidemiologists and public health researchers conducting Mendelian randomisation studies to investigate causal relationships between biomarkers (particularly protein expression) and disease outcomes. However, the paper is primarily a statistical methods contribution rather than an applied agricultural or nutritional health study.

Key measures

Causal effect estimates stability and reliability; comparison of variable selection and estimation methods (stepwise pruning, conditional analysis, principal components analysis, factor analysis, Bayesian variable selection); weak instrument bias assessment

Outcomes reported

The study reviewed and compared statistical methods for conducting two-sample summary-data Mendelian randomisation (MR) analyses using multiple correlated genetic variants from a single gene region, particularly for cis-MR studies using protein expression as a risk factor. The research evaluated method performance through simulation and applied the approaches to case studies examining causal effects of low-density lipoprotein-cholesterol and serum testosterone on coronary heart disease risk.

Theme
Measurement & metrics
Subject
Measurement methods & nutrient profiling
Study type
Research
Study design
Methodological review with simulation study and case studies
Source type
Peer-reviewed study
Status
Published
System type
Laboratory / in vitro
DOI
10.1002/gepi.22506
Catalogue ID
SNmoj1y0b7-lps7f8

Topic tags

Pulse AI · ask about this record

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.