Most Mendelian randomization (MR) of dietary intake use the full set of genome-wide significant (GWS) variants in the instrumental variable (IV), likely biasing causal estimates du
A methodological study demonstrating that GhostKnockoffs with proper Z-score derivation can identify variants with potentially causal effects in related-sample GWASs, validated by simulation and a nine-study meta-analysis, but without direct clinical outcome comparison.